Pre-frontal tDCS improves sustained attention and promotes artificial grammar learning in aphasia: An open-label study

Transcranial direct current stimulation (tDCS) is a potentially ground-breaking form of aphasia treatment, but questions remain related to stimulation targets and underlying cognitive benefits of tDCS. TDCS may be especially useful for increasing cost-effectiveness of aphasia therapy due to its relatively low cost and evidence that it may enhance certain types of therapy outcomes [1,2]. Many studies investigating tDCS in aphasia have used fMRI to locate active regions surrounding participants’ lesions [2,3] then targeted for tDCS stimulation. This individualized approach has been very successful in achieving good response to treatment and is reasonable for a well-funded research study, but is less practical from a clinical perspective as the tools required to precisely locate perilesional targets are expensive and not readily available to most speech-language pathologists.

An alternative approach involves targeting stimulation to areas of the brain that are farther away from the damaged regions, but still involved in language processing. The few studies that have considered this approach in aphasia have primarily focused on stimulating the cerebellum [4,5] or the dorsolateral prefrontal cortex (DLPFC) [6]. Thus far, evidence suggests that non-perilesional tDCS stimulation targets have the potential to improve response to aphasia treatment (e.g., behavioral treatment that focuses on noun/verb retrieval or grammatical sentence production), but more research is needed to better understand the effects of stimulation on different aspects of language.

Artificial grammar training [7] is a paradigm commonly used to study learning of information that requires specific sequential order (e.g., grammar). Studies in healthy controls have shown that tDCS paired with artificial grammar training can result in greater artificial grammar learning [8,9], suggesting that tDCS may also have the potential to enhance artificial grammar learning in aphasia. The objective of this study was to determine if administering anodal transcranial direct current stimulation (tDCS) to the left dorsolateral prefrontal cortex (DLPFC) and cathodal tDCS to right supraorbital region in conjunction with artificial grammar training would result in significant pre- to post-treatment change on a measure of sustained attention and successful artificial grammar learning. This was an open-label study designed to test the feasibility of our treatment protocol in this population and provide preliminary evidence that DLPFC is a reasonable tDCS target for improving attention and artificial grammar learning.

This study was approved by the Syracuse University IRB and all participants provided written consent. Twelve participants (M_age = 63.7 years; 58.3% female) with mild/moderate aphasia (M_QAB[10] = 7.5; 5 Broca’s, 7 Anomic aphasia) were enrolled. Aphasia diagnosis was determined by a speech-language pathologist and left hemisphere stroke was confirmed by radiologist report (M_{time post-onset} = 44.3 months).

All participants received tDCS using 5×5 cm electrodes (anode: left DLPFC [F3]; cathode: right supraorbital region [Fp2]); current was increased to 2.0 mA at the onset of training using a 1×1 tES device (Soterix Medical). This device was pre-programmed by the manufacturer to ramp the current up to the target (2.0 mA in this experiment) over the course of 30 seconds. Once the target current was reached, the timer on the device began counting down each minute of the stimulation session (20 minutes in this experiment) and at the end of the timer countdown, the device automatically ramped the current back down to 0.0 mA over the course of 30 seconds. During each of ten 30-minute training sessions, participants completed 20 minutes of simultaneous tDCS and artificial grammar training and 10 minutes of only artificial grammar training. During artificial grammar training, participants were presented with strings of shapes that followed specific rules of an artificial grammar and were asked to recall them in a matching task using laminated cards while the clinician provided auditory feedback regarding response accuracy.

Sustained attention was measured pre- and post-training using a Continuous Performance Task (CPT) and artificial grammar learning was measured post-training using a 2-choice grammaticality judgment task and an artificial grammar rules test. The CPT, 2-choice grammaticality judgment task, and artificial grammar training protocol were identical to those used during a prior pilot study we conducted in participants without aphasia [9]. The artificial grammar rules test was added to this study and asked participants open-ended questions about the artificial grammar rules they learned in training (e.g., which shapes can occur at the end of a sequence?); participants responded using picture cards.

Nonparametric statistics were used to compare pre- to post-training scores on the CPT and to compare probability-based pre-training scores to post-training scores on the two artificial grammar measures for 11 participants (1 participant withdrew after 5 treatment sessions and did not return for post-testing). Pre-training scores on the artificial grammar learning tasks were estimated based on the probability that a participant would provide an accurate response simply by chance (50% for grammaticality task; 21.5% for artificial grammar rules test). This procedure was used to generate non-zero pre-training scores for each participant using Excel to generate a random number for each participant that fit within the appropriate probability parameters. The post-training standard deviation of each artificial grammar learning task was used to estimate standard deviation.

Participants’ sustained attention accuracy significantly increased following treatment (effect size d = 0.55); participants also demonstrated successful artificial grammar learning by significantly improving their ability to reject ‘ungrammatical’ structures (effect size d = 2.98) and by significantly improving their knowledge of implicitly learned rules of the artificial grammar (effect size d = 11.87). See Table 1 for additional detail. No participants reported adverse effects following tDCS, although most (10 of 12) participants reported feeling a mild burning or itching sensation at the start of each tDCS session.

Table 1.

Results of primary outcome measures.

	Primary Outcome Measures	Testing Session	Mean	Median	Std. Deviation	p-value	Test Statistic	Effect Size (Cohen’s d)
Attention measure	CPTa Accuracy	Pre-Training	322.273	340.000	68.307	**0.005	55.0	0.553
		Post-Training	349.182	350.000	7.960
Artificial Grammar Learning Measures	Accurate Rejection of Ungrammatical Structures	Pre-Training	8.183	8.021	0.984	**0.003	66.0	2.983
		Post-Training	14.455	15.000	2.806
	Grammar Rule Knowledge Accuracy	Pre-Training	2.607	2.437	0.815	**0.003	66.0	11.867
		Post-Training	10.864	11.000	0.552

^**indicates p <.017 (adjusted p-value after Bonferroni correction)

^aCPT = Continuous Performance Test

In the current study, 12 participants with aphasia successfully completed multiple sessions of tDCS/artificial grammar training, and findings suggest that all three primary outcome measures significantly improved following training. Limitations of this study include its small sample size and the lack of a control treatment condition; the open-label design was intended to maximize the number of participants receiving treatment and avoid underpowering the study at this early stage. All participants were able to understand and complete the treatment task, regardless of aphasia type or severity, indicating that the artificial grammar training task is feasible to use with this target clinical population. Results suggest that active tDCS administered to left DLPFC and right supraorbital region may be beneficial for improving sustained attention during language training in persons with aphasia and this improvement may facilitate language learning. Future research should focus on comparing against a control (sham) condition to further investigate the role of tDCS in improving attention and language learning.