Neuromodulation in post-stroke aphasia treatment

Abstract

Purpose of Review

This paper aims to review non-invasive brain stimulation (NIBS) methods to augment speech and language therapy (SLT) for patients with post-stroke aphasia.

Recent Findings

In the past five years there have been more than 30 published studies assessing the effect of transcranial direct current stimulation (tDCS) and transcranial magnetic stimulation (TMS) for improving aphasia in people who have had a stroke. Different approaches to NIBS treatment have been used in post-stroke aphasia treatment including different stimulation locations, stimulation intensity, number of treatment sessions, outcome measures, type of aphasia treatment, and time post-stroke.

Summary

This review of NIBS for post-stroke aphasia shows that both tDCS and TMS can be beneficial for improving speech and language outcomes for patients with stroke. Prior to translating NIBS to clinical practice, further studies are needed to determine optimal tDCS and TMS parameters as well as the mechanisms underlying tDCS and TMS treatment outcomes.

Introduction

Stroke is one of the leading causes of disability worldwide and aphasia or impairment of language is one of the most common consequences. The National Aphasia Association (NAA) estimates that there are more than two million people living with aphasia in the United States, making it more common than Parkinson’s disease, cerebral palsy or muscular dystrophy [1]. Speech and language therapy (SLT) is considered the standard of care for aphasia rehabilitation [2]. A systematic review found evidence of the effectiveness of SLT for people with aphasia following stroke in terms of improved functional communication, reading, writing, and expressive language compared with no therapy [3]. Although the results of the behavioral aphasia treatment studies are encouraging, the effect sizes are somewhat small, which may result in only modest improvements in many cases. Therefore, it is crucial to find ways to improve the effectiveness of aphasia treatment. Non-invasive brain stimulation (NIBS) is one method that has been used to enhance the effects of aphasia treatment. Two of the most common technologies for NIBS are repetitive transcranial magnetic stimulation (rTMS) and transcranial direct current stimulation (tDCS). This article summarizes recent advances in aphasia treatment utilizing NIBS in the last 5 years.

TMS is a non-invasive brain stimulation method that induces changes in neuronal firing via electromagnetic induction. Typically, a figure 8 shaped coil is held over the scalp and used to deliver a changing magnetic field that in turn induces electrical currents in focal regions of the brain [4–7]. Depending on the frequency, duration, and intensity of the stimulation, TMS can lead to depolarization or hyperpolarization of neurons and thus changes in the excitability of stimulated cortex over different time courses [8]. Single pulses of TMS have short-term effects, while repeated pulses of the same intensity have been shown to have more lasting effects [5]. For that reason, repetitive transcranial magnetic stimulation (rTMS) is typically used in therapeutic studies in which long-term benefits are sought. Low frequency rTMS (≤1 Hz) is considered to be inhibitory, while high frequency rTMS (> 1 Hz, and usually ≥5 Hz) is considered to excitatory or facilitatory [9,10]. TMS pulse intensity delivered by the coil during aphasia therapy is typically determined for each participant relative to their individual motor threshold. The motor threshold is usually found by delivering single pulses of TMS to the motor cortex to find the smallest intensity that reliably induces a motor evoked potential in one of the hand muscles [6]. The coil output is adjusted for therapy as a percentage of this motor threshold, most commonly 80%−90%, although threshold and suprathreshold levels may be used as well depending on the goal of stimulation [4–8].

tDCS is a safe and non-invasive approach that is used to modulate cortical excitability [11]. It is usually administered via saline-soaked surface sponge electrodes attached to the scalp and connected to a direct current stimulator with low intensities (1–2 mA) [12]. tDCS can increase or decrease cortical excitability due to a shift of the resting membrane potential of the nerve cells in the brain [13]. Anodal stimulation may lead to depolarization of the neuronal membranes nearest the electric field, and therefore result in greater cortical excitability, whereas cathodal stimulation may lead to hyperpolarization and therefore result in lower cortical excitability. The after-effects of tDCS on cortical excitability are modulated by N-methyl-d-aspartate (NMDA) receptor-dependent processes [14]. The effects of tDCS have been observed up to an hour following a single stimulation session and may persist for days or even months after multiple days of stimulation [15].

Compared with tDCS, rTMS provides the benefits of more focal stimulation with better temporal resolution on the order of milliseconds as opposed to minutes. However, to its detriment, TMS is more expensive; less portable; generates muscle contractions and sound stimuli (a loud audible click) during application that may increase discomfort; carries a risk of inducing seizures, particularly when applied to perilesional areas where hemorrhagic staining may have occurred [16–18]. Additionally, tDCS is more easily paired with simultaneous SLT, making it more amenable to widespread clinical use [19].

TMS studies in Aphasia

Studies investigating the use of rTMS in the treatment of post-stroke aphasia aim to modulate activity in the bilateral language network. Table 1 summarizes a number of such studies published from 2015–2019. Studies for post-stroke aphasia have utilized different approaches to delivering rTMS. The designs of many of these studies are guided by the interhemispheric inhibition hypothesis, which suggests that damaging the language-dominant left hemisphere releases the right hemisphere from transcallosal inhibition. With the loss of left hemisphere transcollosal inhibitory drive, the right hemisphere becomes overactive and in turn sends excess inhibition to the left hemisphere, making recovery of function in the left hemisphere difficult [18, 20–27]. There are two main approaches to rTMS treatment to address this imbalance: inhibitory low frequency rTMS can be applied to the intact right hemisphere to reduce interference, or excitatory high frequency rTMS can be applied to the damaged left hemisphere to stimulate perilesional areas that can be recruited for recovery. Other studies have instead applied excitatory high frequency rTMS to the right hemisphere under the hypothesis that compensatory use of the right hemisphere is not always maladaptive. In addition to the stimulation location and frequency, studies differ in terms of number of treatment sessions (single vs. multiple), outcome measures, use of concurrent speech and language therapy (whether or not therapy is conducted and, if so, which type of treatment), time post stroke [subacute (< 3 months) vs. chronic (> 6 months)], and follow up period [short-term follow-up (e.g. immediately after treatment, 2 weeks post treatment) vs. long-term follow-up (e.g., 6 months post treatment)].

Table 1.

Summary of Reviewed rTMS Studies from 2015–2019.

Author	Participants	Study Design	Stimulation Type	Stimulation Location	Number of Treatment Sessions	Outcome Variables	Results
Hara et al., 2015	50 chronic	Exploratory	inhibitory low frequency rTMS (1Hz)	IFG for non-fluent aphasia, STG for fluent aphasia; stimulate hemisphere contralateral to that identified by fMRI as the compensatory language hemisphere (left or right)	10	SLTA; SPECT	Behavioral language measures correlate with changes in laterality index (LI) of regional cerebral blood flow measured via SPECT. When stimulate right hemisphere: total SLTA score correlated with LI change in BA 44; speaking subscale correlated with LI change in BA 11, 20, 21; writing subscale correlated with LI change in BA 6 and 39. When stimulate left hemisphere: LH stimulation: Speaking subscale correlated with LI change in BA10; reading subscale correlated with LI change in BA 13, 20, 22, 44
Rubi-Fessen et al., 2015	30 subacute	Randomized, sham controlled	inhibitory low frequency rTMS (1Hz)	right IFG	10	Aachen Aphasia Test; Amsterdam-Nijmegen Everyday Language Test; picture naming; functional communication	rTMS led to significant improvement on all measures of linguistic skills and functional communication; 5/10 measures had greater improvement for real vs. sham; generalization to everyday communication observed
Yoon et al., 2015	20 subacute	Randomized, controlled	inhibitory low frequency rTMS (1Hz)	right IFG	20	WAB	Significant improvement in repetition and naming for the rTMS + speech therapy, but not for speech therapy alone
Harvey et al., 2017	9 chronic	Exploratory	inhibitory low frequency rTMS (1Hz)	right IFG	10	Picture naming; semantic fluency; fMRI	Naming improved within rTMS sessions; generalization to naming of novel pictures; naming improvement at 6 month follow-up. On fMRI, posterior shift in activation of right IFG from BA 45 to BA 6, 44, and 46 associated with long-lasting effects; increased left hemisphere recruitment for naming
Santos et al., 2017	13 chronic	Randomized, double blind, sham-controlled, cross over study	inhibitory low frequency rTMS (1Hz); anodal tDCS	tDCS: Left IFG rTMS: Right IFG	1 rTMS, 1 tDCS, 1 sham	Picture naming	No statistically significant difference between A-tDCS, rTMS and sham stimulation in picture naming
Haghighi et al. 2018	12 subacute	Randomized, sham controlled	inhibitory low frequency rTMS (1Hz)	right IFG	10	WAB	Greater improvement for rTMS than sham for AQ, content, fluency, command comprehension, and repetition but not for auditory comprehension and naming
Ren et al. 2019	45 subacute	Randomized, sham controlled	inhibitory low frequency rTMS (1Hz)	right IFG or right posterior STG	15	WAB	rTMS of right IFG group had significantly greater increases in auditory comprehension, repetition, and AQ vs. sham group; rTMS of right STG group had significantly greater increases in repetition, spontaneous speech, and AQ vs. sham group
Rossetti et al. 2019	1 chronic	Case study	inhibitory low frequency rTMS (1Hz)	right IFG	10	Picture naming; Semantic and phonemic fluenc; Stroop test	Phonemic fluency increased immediately after rTMS, and gains continued to 2 month follow-up. Semantic fluency, naming, and Stroop test performance did not improve
Heikkinen et al. 2019	17 chronic	Randomized, sham controlled	inhibitory low frequency rTMS (1Hz)	right IFG	20	WAB; picture naming	Improvement associated with Intensive Language-Action Therapy (ILAT), but no effect of rTMS or effect of combining ILAT and rTMS
Zhang et al. 2017	1 subacute	Case study	excitatory, high frequency rTMS (5Hz)	left IFG	10	WAB; fMRI	WAB scores improved at 2 weeks and 2.5 months post-rTMS. fMRI showed a shift from right hemisphere activation to more focused activation in perilesional left hemisphere areas. Fractional anisotropy increased in left STG
Hara et al. 2017	8 chronic	Exploratory	inhibitory low frequency rTMS (1Hz) or excitatory high frequency rTMS (5Hz)	right IFG; use fNIRS lateralization information to determine whether low or high frequency stimulation group	10	SLTA; fNIRS	Both groups showed significant, comparable improvement on SLTA; fNIRS showed resolution of imbalance of interhemispheric inhibition in low frequency rTMS/left hemisphere activation group and activation of the target hemisphere in the high frequency rTMS /right hemisphere activation group
Hu et al. 2018	40 chronic	Randomized, sham controlled	inhibitory low frequency rTMS (1Hz) or excitatory high frequency rTMS (10Hz)	right IFG	10	WAB	Low frequency rTMS had immediate benefits that persisted long-term. High frequency rTMS had long-term but not immediate benefits. Low frequency rTMS group performed better than high frequency rTMS group immediately after treatment and at 2 months. High frequency group performed better than control group (but not sham group) at 2 months. Placebo effect with better immediate performance by sham than control group immediately after treatment
Griffis et al. (2016)	8 chronic	Exploratory	excitatory iTBS	left IFG	10	Picture naming; Semantic fluency; COWAT; BDAE; functional communication; fMRI	Significant improvement in semantic fluency. fMRI showed increase in left IFG (IFG) activation and decrease in right IFG activation. Reduced connectivity from right to left IFG during covert verb generation. Reduction in right IFG activation was correlated with behavioral improvements in semantic fluency. No gray matter volume changes.
Szaflarski et al. (2018)	12 chronic	Exploratory	excitatory iTBS	left fronto-temporal region	10	WAB; BNT; COWAT; Semantic Fluency; fMRI	Significant improvement on WAB-AQ correlated with decreased BOLD signal in left inferior parietal lobe. Significant improvement in naming correlated with decreased BOLD signal in right IFG. COWAT and SFT did not change significantly but trended toward improvement.
Harvey et al. (2019)	11 chronic	Sham-controlled cross-over design	inhibitatory cTBS	right IFG	1 cTBS, 1 sham	Picture naming	Relative to sham, cTBS improved naming of items that were inconsistently named at baseline but not those that were consistently named incorrectly for those with more severe baseline impairments. Baseline phonological but not semantic naming impairment severity was correlated with improved overall accuracy and with decreased phonological errors.
Georgiou et al. 2019	2 chronic	Case series	inhibitatory cTBS	right IFG	10	BDAE; narrative production; quality of life	Participant 1: improved comprehension and expressive language both post-treatment and at 3 month follow-up and improved overall quality of life. Participant 2: improved comprehension post-treatment that was maintained at 3 month follow-up, but declined on expressive language and overall quality of life
Vuksanović et al. 2015	1 chronic	Case study	excitatory iTBS and inhibitory cTBS	iTBS of left IFG; cTBS of right IFG	15	Picture naming; BDAE; Rey Auditory Verbal Learning Test	Improvements in propositional speech, semantic verbal fluency, auditory comprehension, naming, short-term verbal memory, and verbal learning

Abbreviations: rTMS= repetitive transcranial magnetic stimulation; iTBS= intermittent theta bust stimulation; cTBS= continuous theta burst stimulation; IFG=inferior frontal gyrus; STG=superior temporal gyrus; fMRI=functional magnetic resonance imaging; fNIRS=functional near-infrared spectroscopy; SPECT= Single Photon Emission Computed Tomography; tDCS= transcranial direct current stimulation; WAB= Western Aphasia Battery; AQ=aphasia quotient; BNT= Boston Naming Test; BDAE= Boston Diagnostic Aphasia Examination; BA=Brodmann Area; SLTA=Standard Language Test of Aphasia; COWAT=Controlled Oral Word Association Test

Inhibitory rTMS to right hemisphere regions

The majority of recent studies have utilized the first strategy, inhibiting the right inferior frontal gyrus, the right hemisphere homologue of Broca’s area, with low frequency rTMS. This strategy has led to positive results in terms of language improvement for both subacute [5, 10, 28, 29] and chronic [30–32] post-stroke aphasia. However, the evidence for inhibition of right IFG is not unequivocal: Not all such rTMS studies have reported positive results. For example, Heikkinen and colleagues (2019) [33] reported that improvement in chronic post-stroke aphasia was associated with a specific type of speech and language therapy, Intensive Language-Action Therapy, but that low frequency rTMS did not add to the effect of this language therapy. Santos and colleagues (2017) [16] reported null results when they compared rTMS, tDCS, and sham stimulation in a within-subjects cross-over design; although participants with chronic post-stroke aphasia improved after both tDCS and sham stimulation, there were no significant differences in improvement for the three types of stimulation. Furthermore, there is likely a “file drawer” problem whereby rTMS studies producing null results are left unpublished.

It is also important to note that the right inferior temporal gyrus is not the only viable target for stimulation. Low frequency rTMS of other right hemisphere sites may also be beneficial. For example, Ren and colleagues (2019) [29] reported improvements in language abilities such as repetition and spontaneous speech for participants with subacute stroke who had rTMS of the right superior temporal sulcus relative to the group that received sham stimulation. Determining which right hemisphere sites should be inhibited is an active area of research.

Facilitatory rTMS to the left hemisphere regions

Other studies have attempted the second strategy, facilitating activation of left hemisphere areas. Although this strategy has recently been utilized more frequently in tDCS studies, rTMS studies have utilized this approach as well. For example, Zhang and colleagues (2017) [34] report the results of a case study of a participant with subacute stroke. High frequency rTMS to left inferior frontal gyrus not only led to language improvement as measured by WAB scores, but also to changes in brain activation that were observed with fMRI. Specifically, the participant showed stronger activation in the right hemisphere prior to treatment, but then shifted to more focused activation near damaged tissue in the left hemisphere after treatment.

Facilitatory rTMS to the right hemisphere regions

Alternatively, right hemisphere activation can be seen as compensatory instead of maladaptive [18, 21]. Some have hypothesized that an increase in right hemisphere activation after injury plays an important role over the course of recovery, with early compensatory activation in the right hemisphere being especially beneficial, although it may become maladaptive at more chronic stages of recovery [35–37]. The extent of damage in the left hemisphere language networks may also matter: in those with very large lesions, there is limited perilesional tissue in the left hemisphere that can take over functions and so right hemisphere compensation may be necessary [38, 39]. With this in mind, some rTMS studies have applied high frequency stimulation to the right hemisphere to facilitate compensatory recovery. Hara and colleagues (2017) [40] found positive effects on language not only for inhibitory low frequency rTMS of the right IFG for participants with chronic aphasia who showed left lateralized activation for language prior to treatment, but also for excitatory high frequency rTMS of the right inferior temporal gyrus for those with right lateralized activation for language prior to treatment. That is, those who were already recovering function in the left hemisphere benefited from inhibiting the right hemisphere, whereas those who were compensating with right hemisphere activation benefitted from facilitation of that activation. This suggests that different strategies are appropriate for different types of recovery. Interestingly, it may not be the case that excitatory stimulation of the right hemisphere is only beneficial in a subset of patients. Hu and colleagues (2018) [41] randomized assignment of subacute stroke participants to receive high frequency rTMS of right inferior frontal gyrus, low frequency rTMS of right inferior frontal gyrus, sham stimulation, or standard treatment only as a control. While low frequency stimulation was the most successful both immediately after treatment and at two-month follow-up, there was also long-term improvement relative to controls for high frequency rTMS. That is, both types of stimulation were beneficial. More studies are needed to further probe the conditions under which each type of stimulation is most helpful.

Theta burst stimulation

Recently, interest has grown in using a specific type of rTMS protocol, theta burst stimulation (TBS), in treatment studies. This paradigm was originally designed to mimic the hippocampal firing pattern of rodents in order to induce long-term potentiation [17, 42]. These TBS protocols induce robust, long-lasting changes in activation in much shorter time periods than are necessary for traditional rTMS. Two types of protocols are in use: intermittent TBS (iTBS), which uses a repeating series of short trains of pulses to facilitate activation, and continuous TBS (cTBS), which uses an uninterrupted train of pulses to inhibit activation [43, 44]. Studies of excitatory iTBS of left fronto-temporal regions in chronic stroke have reported behavioral improvements in language that are correlated with reduced BOLD signal in right inferior frontal gyrus on fMRI [17, 44]. That is, this type of stimulation can aid recovery in left hemisphere language regions, inducing a shift from reliance on homologous regions in the right hemisphere. Inhibitory cTBS of the right inferior temporal gyrus is also a promising technique. Case series of cTBS of right IFG alone [43] and cTBS of right inferior frontal gyrus paired with iTBS of left inferior frontal gyrus [45] demonstrated improvement, although not for all domains of language.

tDCS studies in Aphasia

Studies investigating the use of tDCS in the treatment of post-stroke aphasia aim to modulate functional reorganization of language networks by recruiting neurons near the damaged left-hemispheric brain area or by reducing interference with the right-hemispheric language region [46]. Table 2 summarizes a number of such studies published from 2015–2019. Different approaches to tDCS treatment have been used in post-stroke aphasia treatment. This involves excitatory protocols involving anodal stimulation to the left hemisphere, inhibitory protocols involving cathodal stimulation to the right hemisphere, combination of both types of protocols, and right cerebellar tDCS. In addition to the stimulation location, there have differences among studies in terms of stimulation intensity (1mA vs 2 mA), number of treatment sessions (single vs multiple), outcome measures (e.g., naming versus functional communication), type of aphasia treatment (e.g., naming treatment versus auditory comprehension), time post-stroke [subacute (< 3 months) vs chronic ( > 6 months)], and follow up period [short-term follow-up ( e.g., immediately post treatment, 2 weeks post treatment) vs long-term follow-up (e.g., 6 months post treatment)].

Table 2:

Summary of Reviewed tDCS Studies from 2015–2019.

Author	Participants	Study Design	Stimulation Type	Stimulation location	Treatment sessions	Outcome Variable	Results
Branscheidt et al. 2018	16 chronic	Randomized, double blind, sham-controlled, cross over	Anodal tDCS	Left motor cortex	1 session with tDCS and 1 with sham	Lexical Decision	Improved accuracy in lexical decision with anodal tDCS compared to sham.
Fridriksson et al. 2018	74 chronic	Randomized, double blind, sham-controlled Futility design	Anodal tDCS	Left temporal lobe region with the highest naming related activation on the fMRI	15 sessions with either tDCS or sham	Picture Naming	Improved picture naming for anodal tDCS compared to sham.
Pestalozzi et al. 2018	14 chronic	Randomized, double-blind, sham-controlled, crossover	Anodal tDCS	Left dorsolateral prefrontal cortex	1 session with tDCS and 1 with sham	Picture Naming, Verbal Fluency, and Word Repetition	Improved verbal fluency and picture naming speed with anodal tDCS compared to sham.
Spielmann et al. 2018	58 subacute	Randomized, double-blind sham-controlled	Anodal tDCS	Left IFG	5 sessions with tDCS or sham	Picture Naming	No significant improvement in naming with anodal tDCS compared to sham tDCS
Darkow et al. 2017	16 chronic	Randomized, sham-controlled, crossover trial	Anodal tDCS	Left primary motor cortex	1 session with tDCS and 1 with sham	Picture naming, fMRI	No difference in naming accuracy between anodal tDCS and sham. Increased activity in language networks with anodal tDCS compared to sham.
Meinzer et al. 2016	26 chronic	Randomized, double-blind, sham-controlled	Anodal tDCS	Left primary motor cortex	16 sessions (2 sessions per day) with either tDCS or sham	Picture naming and functional communication	Improvement for trained items noted for both anodal tDCS and sham. Generalization to untrained items and functional communication only for anodal tDCS condition
Campana et al. 2015	20 chronic	Randomized, double blind, sham-controlled, cross over	Anodal tDCS	Left IFG	10 sessions with tDCS and 10 sessions with sham	Picture Description, Verb and Noun Naming	Significant improvement in picture description, verb and noun naming with anodal tDCS compared to sham. In addition, damage to distinct left hemispheric structures resulted in lower responses to anodal tDCS.
Wu et al., 2015	12 chronic	Sham-controlled, cross over study	Anodal tDCS	Left posterior perisylvian region	20 sessions with tDCS and 40 with sham	Picture Naming and Auditory Word-Picture Identification, EEG	Improved picture naming and auditory comprehension with anodal tDCS. EEG analysis indicated that improved picture naming correlated with a higher activation level in wide areas of the left hemisphere and in isolated areas of the right hemisphere.
Santos et al. 2017	13 chronic	Randomized, double blind, sham-controlled, cross over	Anodal tDCS and inhibitory low frequency rTMS	tDCS: Left IFG rTMS: Right IFG	1 session with tDCS, 1 session with rTMS and 1 with sham	Picture Naming	No significant difference between A-tDCS, rTMS and sham stimulation in picture naming
da Silva et al. 2018	14 chronic	Randomized, double-blind, sham-controlled	Cathodal tDCS	Right IFG	5 sessions with either tDCS or sham No task (off-line tDCS)	Picture Naming	No significant results were observed for the Snodgrass test. Improved mean time for correct response with strategy for BNT with cathodal tDCS compared to sham
Feil et al. 2019	12 subacute	Randomized, double blind, sham-controlled	Bihemispheric tDCS	Left and Right IFG	10 sessions with either tDCS or sham	Picture Naming, ANELT, AAT	Improved picture naming with bihemispheric tDCS compared to sham immediately after treatment, while ANELT and AAT scores improved at 4 weeks post treatment.
Marangolo et al. 2016	9 chronic	Randomized, double-blind, sham-controlled	Bihemispheric tDCS	Left and right IFG	15 sessions with tDCS and 15 with sham	Word and syllable repetition, fMRI	Improved articulatory accuracy for treated and untreated stimuli for tDCS compared to sham. Furthermore, functional connectivity analysis revealed increased connectivity in the LH with tDCS but only in the RH with sham.
Manenti et al., 2015	1 chronic	Case study	Bihemispheric tDCS	Left and right DLPFC	20 sessions with tDCS	Verb naming	Improved verb naming with treated and untreated items with bihemispheric tDCS. Improvements were maintained at 4, 12, 24, and 48 weeks post treatment.
Sebastian et al. 2017	1 chronic	Case study	Cerebellar tDCS	Right cerebellum	15 sessions with tDCS and 15 sessions with sham	Spelling to Dictation and Written Picture Naming	Improved spelling for trained and untrained words, generalization to written picture naming immediately after and 2 months post-treatment with anodal cerebellar tDCS
Marangolo et al. 2018	12 chronic	Randomized double-blinded, sham controlled, cross over	Cerebellar tDCS	Right Cerebellum	5 cathodal tDCS and 5 sham with verb naming, and 5 cathodal tDCS and 5 sham with verb generation	Verb Naming and Verb Generation	Improved verb generation with cathodal cerebellar tDCS compared to sham.
Shah-Basak et al. 2015	6 chronic	Randomized, sham-controlled, partial cross over	Anodal and Cathodal tDCS	Optimal tDCS montage. Anodal: Left or right frontal lobe, Cathode: Left or right frontal lobe, and Sham	10 sessions with an optimal tDCS montage and 10 with sham	Picture naming	Improved picture naming with left-frontal cathodal-stimulation compared to sham.
Norise et al. 2017	9 chronic	Randomized, sham-controlled, partial cross over	Anodal and cathodal tDCS	Optimal tDCS montage Anodal: Left or right frontal lobe, Cathode: Left or right frontal lobe, and Sham	10 sessions with an optimal tDCS montage and 10 with sham	Language fluency: BDAE Cookie theft	Severe baseline language profile was associated with larger improvements in fluency at the word-level after tDCS but not sham stimulation. These improvements were maintained at the 2-week follow-up.

Abbreviations: tDCS: transcranial direct current stimulation; fMRI: functional magnetic resonance imaging; IFG: inferior frontal gyrus, BNT: Boston Naming Test; LH: Left hemisphere; RH: Right hemisphere; STG: superior temporal gyrus, EEG: electroencephalogram; ANELT: Amsterdam Nijmegen Everyday Language Test; AAT: Aachen Aphasia Test; BDAE: Boston Diagnostic Aphasia Examination

Anodal tDCS over left hemisphere regions (Cortical facilitation of left hemisphere)

The majority of recent studies have utilized this approach with the goal of increasing cortical excitability in the residual tissue within the left hemisphere to facilitate language recovery. Studies have investigated the effect of a single stimulation session [47–49], versus multi session tDCS [50–55]. With respect to stimulation location, studies have used left dorsolateral prefrontal cortex [49], left inferior frontal gyrus [54], left motor cortex [52], left posterior perisylvian region [53], or individualized stimulation location on the basis of functional magnetic resonance imaging (fMRI) [51]. Both single and multi-session tDCS studies have shown positive results in terms of language improvement compared to sham for chronic stroke patients [47, 49, 51–54]. However, there are also studies which did not show any positive effect of tDCS compared to sham [16, 50]. The Santos study [16] was a single session tDCS study and the lack of beneficial effect might indicate the need for more treatment sessions. The study by Spielmann and colleagues [50] failed to show significant effects of tDCS. The authors hypothesized that an effect of tDCS might be difficult to achieve in the subacute phase, as spontaneous recovery is rather high in this phase compared with the relatively stable chronic phase.

It is important to note that most trials have been small; therefore, it is difficult to make strong conclusions about efficacy of left hemisphere stimulation in post-stroke aphasia treatment. The largest tDCS trial so far was done by Fridriksson and colleagues [51] using a futility design, whereby the null hypothesis assumed a benefit for active compared with sham tDCS, while the alternative hypothesis assumed no difference between active and sham tDCS. Utilizing a double-blind, sham controlled trial, 74 individuals with chronic stroke-induced aphasia completed 15 computerized aphasia treatment sessions and were randomized to receive 1 mA A-tDCS or sham tDCS (S-tDCS) to the intact left temporoparietal region for the first 20 min of each session. TDCS was associated with a greater change in number of correctly named pictured objects compared to sham: 13.9 words for tDCS versus 8.2 words for sham. This result did not demonstrate futility with anodal tDCS, suggesting that further evaluation of anodal tDCS is warranted.

Cathodal tDCS over right hemisphere regions (Cortical inhibition of right hemisphere)

The goal of this approach is to reduce the inhibition from the intact right hemisphere in order to boost the recovery process in the left perilesional cortex. This approach is mostly used in rTMS studies as described above, although a few tDCS studies have utilized this approach, including one in the past 5 years [55]. For example, da Silva’s group [55] evaluated the effect of five, 2 mA 20-minute tDCS sessions on naming (Boston and Snodgrass) tasks in 14 chronic stroke patients with aphasia. To induce right hemisphere cortical inhibition a cathode was placed over the cortex homologous to Broca’s area. It should be noted that tDCS was not paired with SLT. No significant difference in results were observed between tDCS and sham group for Snodgrass naming test. The Boston naming test results indicated significant difference between tDCS and sham in one of the variables related to mean time for correct responses with strategy (1.29 seconds for tDCS versus 2. 57 seconds for sham). Unfortunately, it cannot be determined whether pairing tDCS with SLT might have resulted in significant difference for the Snodgrass naming test as well.

Bihemispheric tDCS

As with the previous approach, only a few studies have investigated the effect of bihemispheric tDCS stimulation to augment aphasia treatment [56–58]. The goal of this approach is to deliver tDCS over both hemispheres, aiming at concomitantly increasing left hemisphere cortical excitability with anodal stimulation and decreasing right hemisphere excitability with cathodal stimulation. Initial studies of bihemispheric tDCS effects in chronic aphasia found significant improvement in language skills when combined with behavioral treatment compared to sham stimulation [57, 58]. Furthermore, functional connectivity analysis using resting state fMRI showed that connectivity changes after the sham condition were confined to the right brain hemisphere regions, whereas real tDCS stimulation yielded stronger functional connectivity increase in the left hemisphere regions [57]. In a recent study on subacute stroke patients with aphasia [56], 10 sessions of bihemispheric tDCS (anode on left inferior frontal gyrus (IFG) and cathode on right IFG) combined with SLT significantly improved picture naming in the tDCS group but not in the sham group. Improved picture naming induced by tDCS could be due to an early left shift of language-associated functions.

Right Cerebellar tDCS

The newest approach to neuromodulation for aphasia is targeting the right cerebellum. The right posterolateral cerebellum is involved in various aspects of cognitive and language processing [59]. In addition, the right cerebellum is thought to be particularly important during skill learning, contributing to the optimization and automatization of performance; therefore, it is possible that cerebellar neuromodulation – acting via the established cerebro-cerebellar circuits with left-hemisphere language regions – could enhance language recovery post-stroke [60]. So far, results from two cerebellar tDCS studies show positive results on language improvement [61, 62]. In a patient with a large bilateral frontoparietal and insular infarct, Sebastian et al. [61] found that both anodal and sham stimulation coupled with SLT resulted in improved spelling to dictation for trained and untrained words immediately after and 2 months post-treatment. However, improvement was greater with anodal tDCS than with sham, especially for untrained items. In addition, generalization to written picture naming was only observed with tDCS. Another study by Marangolo et al. [62] explored the effect of cerebellar tDCS coupled with language treatment for verb improvement in 12 patients with chronic aphasia. The results showed a significant improvement only after cathodal stimulation in the verb generation task.

It is still unclear which tDCS stimulation approach is optimal for facilitation of aphasia recovery, although the majority of research groups have chosen to investigate stimulation of the left hemisphere. Hamilton and colleagues [63, 64] have used an innovative approach to figure out optimal stimulation montage for individual with post-stroke aphasia. In a two-phase study, participants underwent anodal- and cathodal-tDCS of the left and right frontal areas, as well as a sham condition, in separate sessions. A preferred electrode montage was established for each participant by assessing transient improvement on a picture-naming task. In Phase 2—a randomized, sham-controlled, partial cross over design, tDCS was administered over 10-days using the individualized optimal electrode configuration that was identified in Phase 1. Results showed that naming improvement was most pronounced after left-frontal cathodal-stimulation compared to sham stimulation [64]. In a follow up study, they found that severe baseline language profile was associated with larger improvements in fluency at the word-level after tDCS but not sham stimulation [63]. Larger randomized control trials are needed to answer the question regarding optimal stimulation location. So far, only two tDCS studies have included participants with more than 50 patients [50, 51].

Conclusion and future directions

The results of the studies reviewed here indicate that TMS and tDCS are promising tools for the treatment of aphasia. However, wide variation in experimental factors including different types of aphasia, lesion site and location, tDCS/TMS stimulation parameters, different types of SLT combined with treatment, therapy duration, and different outcome measures, presents a major challenge to interpreting the findings. Many questions need to be answered before we can make strong conclusions about the efficacy of these neuromodulation tools in aphasia treatment. One of the most important questions pertains to determining the optimal stimulation location. It remains unclear which area of the brain (left, right, or cerebellum), and which kind of stimulation (inhibitory or excitatory) are more effective in augmenting aphasia treatment. Variability in lesion location and size as well as inter-individual differences in brain reorganization post-stroke could further complicate efforts to determining the optimal stimulation location. The majority of recent tDCS studies have focused on stimulating the left perilesional cortex, whereas majority of the rTMS studies have focused on inhibiting the right hemisphere homologue of Broca’s area. Right cerebellar stimulation could potentially serve as a single target location that could be used across patients with aphasia with varying site and size of lesion in the left hemisphere [60]. Further, cerebellar tDCS could potentially avoid the extra cost and time that associated with procedures such as fMRI to identify perilesional stimulation targets on an individual basis [51].

Other open questions relate to which individuals are likely to benefit from NIBS and which domains of language are likely to improve. Additional studies are needed to develop biomarkers and language profiles that have the potential to distinguish patient subgroups and thus identify persons who are likely to respond favorably to NIBS [65]. A study by Fridriksson and colleagues [66] utilizing data from their 2018 trial investigated whether aphasia treatment outcome is influenced by interaction between anodal-tDCS and a single-nucleotide polymorphism of the brain-derived neurotrophic factor (BDNF) gene. Stroke patients with the normal val/val genotype who received tDCS showed greater response to aphasia treatment than val/val participants who received sham, and greater response to aphasia treatment than the Met allele carriers, regardless of tDCS condition. Similarly, Harvey and colleagues [23], investigated which individuals are likely to benefit from inhibitory cTBS of the right inferior frontal gyrus pars triangularis. After characterizing the nature of their chronic stroke participants’ naming deficits, they asked participants to name items before and after cTBS and sham stimulation sessions. They found that those with the most severe baseline deficits did improve after only one session of cTBS, becoming more likely to correctly name pictures that they had inconsistently named prior to treatment. Furthermore, phonological impairment severity at baseline was correlated with improvements in overall accuracy of naming and with decreases in phonological errors. This was not the case for semantic impairment severity. These results suggest that cTBS enhanced naming by facilitating phonological retrieval, and that individuals with this type of naming impairment could benefit the most from this type of stimulation.

Future studies should also focus on the efficacy of NIBS in functional communication skills. Most of the studies to date have focused on naming, but improvement in naming may not always be followed by an improvement in functional communication or quality of life. For example, a recent Cochrane review [67] found no evidence of the effectiveness of tDCS (anodal tDCS, cathodal tDCS and Dual-tDCS) versus control (sham tDCS) for improving functional communication in people with post-stroke aphasia (low quality of evidence). The study found limited evidence that tDCS may improve naming performance in naming nouns (moderate quality of evidence), but not for verbs (very low quality of evidence) at the end of the intervention period and possibly also at follow-up. Greater understanding of the optimal tDCS and rTMS parameters to enhance aphasia recovery in individual patients with different lesion location/size and language profiles may one day assist with the translation of NIBS protocols into everyday clinical practice.