Thirty-two consecutive patients (22 males and 10 females; mean age 62.1 years; range, 39–82 years) admitted for rehabilitation at the department of rehabilitation in Asan Medical Center between October 2013 and December 2014 were recruited according to the following criteria: 1) first ever stroke, 2) left visuospatial neglect, defined as >6.33 mm average deviation from the center line on the line bisection test (LBT) [11], and 3) diagnosed as right cerebral ischemic or hemorrhagic stroke. Patients were excluded if they met any of the following criteria: 1) severe cognitive dysfunction or aphasia, 2) contraindications for tDCS, such as history of previous seizure, major head trauma, previous brain operation, a metal implant in the brain, or a pacemaker, or 3) systemic disease or ongoing neoplasia. Of the 32 consecutive patients who met the criteria in the recruitment period, 2 were lost to follow-up because of early discharge. Thus, 30 patients were analyzed in this study. All participants provided written informed consent, and our local Ethics Committee approved the study protocol.

Patients were identified by a number assigned by a centralized computer-generated randomization code. The patients were randomly allocated to 1 of the following 3 groups: 1) the anodal group, 2) the cathodal group, or 3) the sham group. The anodal group received anodal tDCS over the right PPC, the cathodal group received cathodal tDCS over the left PPC, and the sham group received sham tDCS in the same way as active stimulation, but the stimulator was turned off after 30 seconds.

Each patient underwent 15 sessions of tDCS. Sessions were 5 times per week for 3 weeks. A direct current was delivered by 2 sets of battery-powered devices named Phoresor II Auto Model PM850 (IOMED Inc., Salt Lake City, UT, USA) using 2 pairs of saline-soaked sponge electrodes (5 cm×5 cm). Stimulation was delivered while the patient was receiving conventional occupational therapy. All patients received conventional physical therapy throughout the duration of the 3 week tDCS protocol. For the real stimulation, a constant current of 2 mA was delivered for 30 minutes. Sham stimulation was performed in the same way as active stimulation, but the stimulator was turned off after 30 seconds. This ensured that subjects could feel the initial itching sensation at the beginning of tDCS and allowed for a successful blinding of the subjects to their stimulation condition [12]. The tDCS stimulation site corresponded with position P3, which is localized over the left PPC, for the cathodal group; and position P4, which is localized over the right PPC, for the anodal group. Sham tDCS was performed in the same way as for anodal group stimulation, with the stimulation site at P4 and reference electrode over Cz, but the stimulator was turned off after 30 seconds.

These positions are according to the 10/20 electroencephalography system. These locations were previously shown to overlie the PPC in close proximity to the intraparietal sulcus [11]. The reference electrode was placed over Cz. The choice of Cz was based on previous studies that investigated the effect of tDCS on the primary visual cortex [13] and parieto-temporal areas. There was no difference in the total amount of rehabilitation time between the 3 groups.

Each outcome measure was evaluated before the treatment and 1 week after completing the 3 weeks of treatment. Severity of neglect was assessed with the motor-free visual perception test (MVPT), the LBT [14], the star cancellation test (SCT), and the Catherine Bergego Scale (CBS) [15].

The MVPT is a 36-item multiple-choice test that evaluates 5 sub-dimensions of visuospatial neglect: visual discrimination, figure-ground discrimination, spatial relationship, visual closure, and visual memory. The parameters that quantify left response behavior (left or right response behavior, raw score, left or right performance behavior, and visual perception processing time) were used. The total score of left response behavior ranges from 0 to 21, and lower scores indicate more severe visuospatial neglect.

The LBT was performed using the method of Schenkenberg et al. [14]. Twenty lines were drawn on an A4-size sheet of white paper, parallel to the long axis. Eighteen of these lines were organized into 3 sets of 6: 1 set lay primarily on the left of the paper, 1 set lay at the center of the paper, and 1 set lay primarily on the right of the paper. Patients were asked to mark the center points of each of the 18 lines. The deviation from the patients' mark and the true line center was measured for each line, and averaged across all 18 lines for each patient.

In the SCT, 56 stars were mixed with other symbols and presented on an A4-size sheet of white paper. The left and right half of the paper each contained 27 stars, 2 stars were at the midline, and patients were asked to mark all the stars. The number of stars marked on the left side was counted. The score ranged from 0 to 27, with lower scores indicating more severe visuospatial neglect.

The CBS [15] is a questionnaire based on a direct observation of the patient's functioning by the situations re-created in the clinic in 10 real-life situations, such as grooming, dressing, and wheelchair driving. The score ranges from 0 to 30, with high scores indicating more severe visuospatial neglect.

The Korean version of the Modified Barthel Index (K-MBI) was used to evaluate activities of daily living function. The K-MBI evaluates 10 areas of functioning, and scores indicate the degree of independence. It is scored from 0 (total dependence) to 100 (independence), depending on the level of assistance needed.

Cognitive impairment was assessed using the Korean version of the Mini-Mental State Examination (K-MMSE), and basic motor skills were assessed using Functional Ambulation Classification (FAC).

Fisher exact test and the chi-square test were used to compare demographic and baseline clinical variables in the 3 groups (anodal, cathodal, and sham). Change in MVPT, LBT, SCT, CBS, K-MBI, and FAC from initial evaluation to the follow-up evaluation was evaluated using the Wilcoxon signed-rank test. The change in MVPT, LBT, SCT, CBS, K-MBI, and FAC from initial evaluation to follow-up evaluation was compared between the 3 groups using the Kruskal–Wallis test. When significant, the Mann–Whitney U test was performed for pairwise comparison (anodal vs. sham, cathodal vs. sham, and anodal vs. cathodal).

All statistical analyses were performed using SPSS ver. 18.0 for Windows (SPSS Inc., Chicago, IL, USA). A p-value <0.05 was considered significant for the Wilcoxon signed-rank test and the Kruskal–Wallis test, and a p-value <0.017 was considered significant for the Mann–Whitney U test.