Edited by: David Copland, University of Queensland, Australia
Reviewed by: Giancarlo Zito, Fatebenefratelli Hospital, Italy; Svetlana Malyutina, National Research University Higher School of Economics, Russia
*Correspondence: Argye E. Hillis
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
People with post-stroke aphasia may have some degree of chronic deficit for which current rehabilitative treatments are variably effective. Accumulating evidence suggests that transcranial direct current stimulation (tDCS) may be useful for enhancing the effects of behavioral aphasia treatment. However, it remains unclear which brain regions should be stimulated to optimize effects on language recovery. Here, we report on the therapeutic potential of right cerebellar tDCS in augmenting language recovery in SMY, who sustained bilateral MCA infarct resulting in aphasia and anarthria. We investigated the effects of 15 sessions of anodal cerebellar tDCS coupled with spelling therapy using a randomized, double-blind, sham controlled within-subject crossover trial. We also investigated changes in functional connectivity using resting state functional magnetic resonance imaging before and 2 months post-treatment. Both anodal and sham treatments resulted in improved spelling to dictation for trained and untrained words immediately after and 2 months post-treatment. However, there was greater improvement with tDCS than with sham, especially for untrained words. Further, generalization to written picture naming was only noted during tDCS but not with sham. The resting state functional connectivity data indicate that improvement in spelling was accompanied by an increase in cerebro-cerebellar network connectivity. These results highlight the therapeutic potential of right cerebellar tDCS to augment spelling therapy in an individual with large bilateral chronic strokes.
Aphasia is a leading cause of disability following stroke and can affect every aspect of daily life, including interpersonal relationships, work, and community interactions. Speech-language therapy is the mainstay of treatment. Therapy is beneficial for language recovery; however, gains in therapy are variable and progress may be slow, especially after large, chronic left hemisphere lesions (Brady et al.,
This case study illustrates the potential usefulness of a novel electrode placement for tDCS augmentation of language therapy in chronic post-stroke aphasia: the right cerebellum.
Evidence from functional neuroimaging and neuroanatomical investigations indicate that the right cerebellum is important for language and cognitive functions (e.g., Leiner et al.,
Given the role of the cognitive and language functions of the cerebellum and the ability of cerebellar tDCS to modify behavior in healthy individuals, cerebellar tDCS may have a uniquely valuable therapeutic role for individuals with aphasia. Furthermore, cerebellum can be stimulated even in patients with aphasia associated with bilateral hemispheric strokes. In addition, the cerebellum is regarded as an important region involved in skill learning (Morton and Bastian,
Here, we report behavioral and neural effects of right cerebellar tDCS with behavioral spelling treatment in a participant who sustained bilateral MCA infarct resulting in aphasia and complete anarthria. Participant SMY is mute following his second stroke but has retained some ability to write and type. Because he depends on writing to communicate, recognizable spelling is critical for effective social function. Therefore, cerebellar tDCS plus behavioral spelling treatment could improve spelling recovery through its roles in language and learning. We sought to evaluate the following hypotheses: (1) Improvement in spelling to dictation (in treated and untreated words) will be greater with tDCS + spelling treatment than with sham + spelling treatment; (2) Improvement will last longer after tDCS treatment than sham treatment at 2 months post-treatment; (3) Improvement in other language tasks (written picture naming) will be greater after tDCS than sham; (4) Functional connectivity between the right cerebellum and the residual left and right hemisphere language regions of interest will be greater post-treatment compared to pre-treatment.
SMY is a 57-year-old, right-handed man with a master's degree, employed as an architect until he had an ischemic stroke due to carotid dissection. MRI revealed left MCA territory infarct, involving frontal, temporal, and insular cortex. This stroke resulted in right hemiparesis and aphasia. He underwent extensive inpatient and outpatient rehabilitation, and showed resolution of hemiparesis and substantial improvement in language. He survived a second (right hemisphere) stroke due to carotid dissection 4 years later. MRI revealed acute infarct of right MCA territory, involving the fronto-parietal and insular cortex (Figure
He was enrolled in the study in 2015, 5 years after his second stroke. At the time of his enrollment, SMY was independent in activities of daily living and resided with his wife. SMY scored 26/28 on the auditory comprehension subtest on the Aphasia Diagnostic Profile test (Helm-Estabrooks,
Given that SMY is mute and the focus of treatment was on spelling, only written language was assessed in detail. SMY's narrative writing consisted of simple sentences with frequent phonologically implausible nonword errors (center → cect), semantic errors (garage → house), and letter omissions (piano → pian). SMY was able to write simple and common 3-letter words and some 4-letter words without difficulty. He was not able to identify his errors in writing. The Johns Hopkins Dysgraphia Battery (Goodman and Caramazza,
Spelling to dictation of words and pseudowords was very impaired (12% accurate on words and 0% accurate on pseudowords). On words, he showed a significant effect of grammatical word class (e.g., verbs vs. nouns 36 vs. 21% accurate; χ2 = 4.8;
Errors in spelling to dictation were mostly phonologically implausible nonword errors (e.g., parent → parpe) and some unrelated word errors (palace → pea). His 0% accuracy on nonword errors indicates impairment at the level of sublexical mechanisms for phoneme-grapheme conversion (PGC), although he was also impaired in access to written word forms. His word length effect can be explained by greater opportunity to err on longer words in his attempts to rely on (impaired) PGC. We decided to treat sublexical spelling—PGC—as a first step, to give him some rules he could rely on to at least produce a plausible spelling.
SMY's written picture naming performance was assessed using the Philadelphia Naming Test (PNT, Roach et al.,
We used a double-blind, within-subject crossover trial design, with random order of treatments. There were two experimental conditions: “right cerebellar tDCS + behavioral (spelling) treatment” and “sham tDCS + behavioral treatment.” Each condition consisted of 15 consecutive training sessions, 3–5 per week, separated by 2 months. Evaluation took place before, immediately after, and 2 months post-treatment for each condition (Tsapkini et al.,
We employed a spelling treatment protocol previously described (Tsapkini et al.,
Trained words | 0/40 | 21/40 ( |
13/40 ( |
13/40 | 39/40 ( |
39/40 ( |
Untrained words | 0/40 | 11/40 ( |
11/40 ( |
11/40 | 33/40 ( |
36/40 ( |
Written naming | 121/175 | 121/175 ( |
122/175 ( |
122/175 | 143/175 ( |
145/175 ( |
The behavioral spelling treatment consisted of training PGC in the context of each dictated word practiced. For each trained word, SMY was asked to point to the letter corresponding to each phoneme (from a set of letters). If he was correct, he was reinforced. If he was incorrect, the clinician pointed to the correct letter. Then SMY was asked to write the letter(s) corresponding to a particular phoneme for the trained word. If he was incorrect, the clinician wrote the correct letter. Finally, SMY was explicitly instructed in PGC for all letter-sounds of the word. Each session consisted of teaching the PGC of five trained words; when SMY met criteria (90% accuracy across three sessions), new words were introduced. It should be noted that SMY did not receive any other therapy except for support groups, including during follow-up periods.
To understand the electric field distribution of right cerebellar tDCS, we completed a modeling study. One high resolution T1 MRI-scan (1 mm3 voxels) of a healthy control extending between the c7 vertebra and the vertex was segmented into 11 tissue compartments using automated algorithm and manual segmentation techniques using ScanIP (Simpleware) as previously described (Datta et al.,
Resting state fMRI was acquired twice: prior to the start of the study and 2 months after the completion of the study (6 month interval between scans). Scans were acquired on a 3-T Philips Achieva MRI scanner with a 32-channel head coil. Resting state images were acquired using EPI and the following scan parameters: TR = 2000 ms, TE = 30 ms, flip-angle = 90, matrix = 64 × 64, FOV = 240 × 240 mm, 35 3 mm parallel axial slices covering the whole brain, 210 volumes. High resolution 3D MPRAGE was acquired in the sagittal plan utilizing a multishot, turbo field echo pulse sequence and the following scan parameters: TR = 6800 ms, TE = 31 ms, matrix = 256 × 256, FOV = 256 × 240 × 240 mm, 170 1 mm slices covering the whole brain. One normal control (59/Female) was scanned longitudinally using the same scan interval as SMY.
Changes in connectivity were examined between the right cerebellum and peri-lesional language regions of interest (ROIs) in the LH and RH. Only non-lesioned regions were included in the ROI. The ROI included: superior frontal gyrus (SFG), superior frontal gyrus_prefrontal cortex (SFG_PFC), middle frontal gyrus/dorsolateral prefrontal cortex (MFG_DLPC), middle temporal gyrus pole (MTG_pole), inferior temporal gyrus (ITG), fusiform gyrus (FG), left and right cerebellum. For a full description of resting state-processing steps see Sebastian et al. (
SMY showed an overall improvement in spelling with a notable increase in writing speed. Trained words were the same for the sham and tDCS conditions. However, out of the 40 trained words, SMY reached criterion for only 17 trained words during the sham condition. On an average, he required 4.5 (range: 3–7) sessions to reach criterion for each trained word. SMY reached criterion more quickly during the tDCS condition. He was able to reach criterion in 3.2 sessions (range: 3–4).
SMY correctly spelled to dictation 21/40 trained words and 11/40 untrained words after sham treatment, whereas he correctly spelled to dictation 39/40 trained words and 33/40 untrained words after tDCS treatment. Therefore, both sham and tDCS treatments were effective for trained and untrained words immediately post-treatment, but there was significantly greater improvement with tDCS than with sham for both trained and untrained words (trained words: χ2 = 14.77,
SMY showed a notable change in spelling error types post-treatment, especially after the tDCS condition. Before treatment, SMY's spelling to dictation was 0% correct (40/40 errors) for both trained and untrained words (trained words: 33/40 or 82.5% phonologically implausible nonword errors; 3/40 or 7.5% phonologically plausible nonword errors; 4/40 or 10% unrelated word errors, untrained words: 35/40 or 87.5% phonologically implausible nonword errors; 2/40 or 5% phonologically plausible nonword errors; 3/40 or 7.5% unrelated word errors). After sham treatment, SMY made 19 errors on the 40 trained words (15/19 or 79% phonologically implausible nonword errors; 1/19 or 5% phonologically plausible nonword errors; 3/19 or 16% unrelated word errors) and 29 errors on the 40 untrained words (26/29 or 90% phonologically implausible nonword errors; 1/29 or 3% phonologically plausible nonword errors; 2/29 or 7% unrelated word errors). After tDCS treatment, he made one error on the trained words and seven errors on the untrained words. All errors were phonologically plausible nonword errors (e.g., pigeon → peigon), and were thus more functional than pretreatment errors.
Resting state functional connectivity analysis data for SMY and the normal control are shown in Figure
This report is, to our knowledge, the first to use cerebellar neuromodulation to augment spelling recovery in an individual with large bilateral chronic strokes. This case study illustrated the potential usefulness of novel electrode placement for tDCS augmentation of language therapy in chronic post-stroke aphasia. Results suggest that anodal tDCS of the right cerebellum coupled with behavioral therapy is more effective than behavioral therapy alone in improving spelling to dictation. Furthermore, generalization to written picture naming was facilitated by tDCS, not sham. Finally, the resting state functional connectivity data indicate that improvement in spelling is accompanied by an increase in cerebro-cerebellar network connectivity.
We found robust improvement in SMY's spelling skills for both trained and untrained words especially after tDCS. Robust effects could be due to the duration and intensity of treatment (15 sessions, 3–5 per week). Alternatively, long-term stimulation may have induced long-term potentiation of neurons that may have lowered the threshold of neuronal excitability and subsequent modification in synaptic connectivity in the areas applied (e.g., Fritsch et al.,
Although, right cerebellum is not traditionally considered to be associated with spelling, we show that cerebellar tDCS along with spelling therapy resulted in significant improvement in SMY's spelling abilities. Given that SMY sustained two large strokes involving the left and right hemispheres, tDCS may have enhanced changes in neuroplasticity resulting in modification of networks underlying spelling.
This study, although preliminary, yielded interesting findings and promising avenues for further studies on cerebellar tDCS in post-stroke aphasia. Limitations include that this was a case study. In addition, tDCS followed sham in this experimental design and thus any extra benefits of tDCS might in fact be the benefits of having a second treatment period after already having had the first treatment period and time for consolidation. Cerebellar tDCS may not have comparable beneficial effects in patients with unilateral LH lesions. Also, resting state data were not acquired before and after completion of each treatment condition (i.e., sham and tDCS). Therefore, it is not possible to disentangle the network connectivity changes associated with tDCS vs. behavioral treatment.
All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication. AH, RS, KT, and PC were involved in conception and design, analysis and interpretation of data. RS, SS, CL, AM, CD, and DT performed language evaluation, tDCS treatment, and or neuroimaging data acquisition. AM and MB performed computational modeling and interpretation. AF performed neuroimaging analysis and interpretation. RS, AH, KT, DT, and PC drafted the article and revised it critically for important intellectual content.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We would like to thank Dr. Julius Fridriksson, Dr. Chris Rorden, and Taylor Hanayik for providing the tDCS device and computer related assistance. This research was supported by NIDCD K99 DC015554, R01 DC05375, and P50 DC014664. The content is solely the responsibility of the authors and does not necessarily represent the views of the National Institutes of Health.
The Supplementary Material for this article can be found online at: