Over the past 10 years, transcranial direct current stimulation (tDCS) has been widely used to modulate cortical excitability to the benefit of cognitive and motor functions, in both healthy and clinical populations ( 1 ). However, studies investigating the effects of tDCS on control over eye movements have been scarce. Only recently evidence has emerged demonstrating that positively charged anodal tDCS applied over the frontal eye field (FEF) can be used to improve saccadic eye movement control in healthy young adults ( 2 ). Given that healthy aging and a large number of age-related clinical conditions (e.g., mild cognitive impairment and Alzheimer’s and Parkinson’s disease) are associated with reduced control over the eye movement system, particularly when a high level of strategic control is required ( 36 ), findings in Kanai, Muggleton ( 2 ) are exciting as they suggest that anodal tDCS may be a useful therapeutic tool for improving voluntary control over the oculomotor system in impaired populations.

Voluntary control over saccadic eye movements involves complex underlying neuromechanisms by which cortical oculomotor regions must be able to impose topdown regulation over subcortical oculomotor regions ( 7 ). The antisaccade paradigm ( 8 ) is a tool commonly used for behavioral measurement of voluntary control over saccadic eye movements. A successful antisaccade involves moving the eyes in the opposite direction when a stimulus suddenly appears in the peripheral visual field. This capability involves two control processes: 1) suppressing an unwanted reflexive prosaccade toward the peripheral stimulus; 2) voluntarily generating an eye movement away from the peripheral stimulus to the mirror position ( 9, 10 ). The frontal subregions most commonly posited to underpin accurate performance of antisaccades are the FEF and dorsolateral prefrontal cortex (DLPFC; 11, 12). However, as reviewed in Chen and Machado ( 13 ), the contributions of the FEF and DLPFC to the suppression of reflexive prosaccades and the generation of correct antisaccades remain unclear. While the literature is in agreement that the FEF is the key region supporting generation of volitional eye movements, there is disagreement as to which frontal subregion supports suppression of reflexive eye movements. Specifically, some have reported evidence implicating the FEF as the key region supporting suppression of reflexive eye movements ( 14 ), while others have claimed that DLPFC (Brodmann’s area 46) plays the main role in suppressing reflexive eye movements, as reviewed in PierrotDeseilligny, Milea ( 15 ).

The one study ( 2 ) that assessed the influences of tDCS over cerebral cortex on oculomotor behavior found that in healthy young adults anodal tDCS over the FEF influenced subsequent antisaccade performance such that reflexive errors were reduced contralaterally without any effect on correct antisaccade latencies, and in addition subsequent correct prosaccade latencies were shortened contralaterally. These findings indicate that while anodal tDCS over the FEF facilitates suppression of unwanted contraversive reflexive eye movements, it also speeds the latencies of wanted contraversive reflexive eye movements. These anodal tDCS benefits peaked 10 to 30 minutes post stimulation. In this seminal study, electrode positioning over the FEF was determined based on predefined standardized coordinates using structural magnetic resonance imaging (MRI) of each individual. In the current study, we tested whether the benefits reported in Kanai, Muggleton ( 2 ) can be induced using a more clinically practical protocol that does not entail expensive tools or time consuming procedures (e.g., MRI) to determine electrode positioning, as using such tools falls outside available resources in many clinical settings.

In addition to assessing anodal tDCS over the FEF, in the current study we assessed whether applying anodal tDCS over DLPFC might also benefit saccadic eye movement control, especially with respect to suppressing unwanted reflexive saccades, as one might predict based on human brain lesion studies ( 16-18 ). Furthermore, in the current study, we assessed whether saccadic eye movement control benefits extend to older adults. Ample evidence from non-oculomotor studies indicates tDCS confers more robust cognitive benefits in older adults ( 19 ), presumably due to far more room for improvement and thus greater potential for benefit. However, no studies to date have assessed whether the same applies to saccadic eye movement control. In testing the efficacy of tDCS to improve saccadic eye movement control, we compared oculomotor behavior ipsilateral versus contralateral to the anodal electrode (as per 2) and we also included a sham control condition (in contrast to 2). This enabled us to determine whether performance contralateral to the FEF and DLPFC electrodes was superior to ipsilateral performance, and also whether it was superior to performance contralateral to sham stimulation.

In sum, the purpose of the current study was threefold: 1) determine whether benefits of anodal tDCS on saccadic eye movement behavior can be induced using a clinically practical protocol; 2) determine whether anodal tDCS over DLPFC also benefits saccadic eye movement control; 3) determine whether benefits extend to older adults.

To determine if performance varied across the blocks, initial repeated-measures analyses of variance (ANOVAs), with stimulation condition, saccade direction, and saccade block as factors, were performed for each of the measured variables of interest (prosaccade latencies, antisaccade latencies, and reflexive error rates during antisaccades). The results revealed no main effect of block for the latency variables, but there was a main effect of block for reflexive error rates during in responsiveness to brain stimulation ( 34 ), to determine whether a subset of the participants benefitted from active stimulation, each individual’s data was checked for any apparent asymmetries in the active stimulation conditions consistent with superior performance contralateral versus ipsilateral, and if so contralateral active versus sham; paired-samples t tests tested whether any of the differences reached significance. Table 1 summarizes the mean of the median correct response latencies and reflexive error rates during antisaccades for each stimulation condition in each age group. Tables S1 and antisaccades, F(2, 58) = 9.398, p < .001, r = .495, reflecting increasing reflexive error rates across blocks, presumably due to fatigue. However, since saccade block did not interact with stimulation condition or saccade direction for any of the measured variables of interest (all ps > .200), the data were collapsed across blocks in the mixed ANOVAs reported below, all of which included age group as a between-participant factor, and stimulation condition and saccade direction as within-participant factors. Regardless of the ANOVA results, paired samples t tests assessed hemispheric asymmetries in the active conditions and differences against sham stimulation. In addition, in light of individual differences S2 (in Appendix) detail the results of each group-level t test.

Using a more clinically practical protocol, the current study tested whether anodal tDCS over the FEF can induce oculomotor benefits similar to those reported in young adults in Kanai, Muggleton ( 2 ), and in addition assessed whether applying anodal tDCS over DLPFC might also benefit oculomotor behavior and whether these benefits extend to older adults, who are known to have saccadic eye movement control deficits ( 13 ). Overall the results revealed no evidence of oculomotor benefits following anodal tDCS, despite the sample size in the current study exceeding that used in Kanai, Muggleton ( 2 ). Specifically, group analyses showed no differences in the active stimulation conditions relative to sham stimulation, and an asymmetry in saccadic eye movement behavior arose only in the active DLPFC stimulation condition (for reflexive errors in the full mixed-age sample), but this did not reflect better performance relative to sham performance, and the sham condition showed a similar pattern. Analyses of individual participants backed up the null results at the group level, with significant effects relative to sham stimulation occurring in less than 5% of the participants (which is consistent with chance levels, as alpha was set to .05). These results indicate that neither active stimulation site (FEF or DLPFC) afforded better saccadic eye movement control. The absence of oculomotor benefits arose in both age groups, despite the older adults exhibiting the expected saccadic eye movement control deficits that indicate ample room for improvement. These negative outcomes indicate that the clinically practical protocol utilized in the current study was ineffective.

One of the main findings reported in Kanai, Muggleton ( 2 ) was that anodal tDCS over the FEF reduced reflexive error rates toward contralateral relative to ipsilateral antisaccade signals. This pattern was also demonstrated in the current study, although the asymmetry did not reach significance. However, as shown in Figure 4, the same pattern also occurred in the sham stimulation condition. Given that Kanai, Muggleton ( 2 ) did not include a sham stimulation condition, it is not possible to determine whether the lower rate of contralateral versus ipsilateral reflexive errors occurred as a result of the tDCS. To determine this, one would need to replicate the protocol used in Kanai, Muggleton ( 2 ) with the addition of a sham stimulation condition. The fact that the current study showed similar asymmetric patterns in the active and sham conditions highlights the need for a sham control comparison condition to confirm whether any observed asymmetries are specifically attributable to tDCS. The other main finding reported in Kanai, Muggleton ( 2 ) is that anodal tDCS over the FEF shortened prosaccade latencies contralateral versus ipsilateral to the stimulated hemisphere. This pattern was not demonstrated in the current study in either age group (see Figure 2). Furthermore, none of the older adults and only three of the 20 young adults showed this pattern, and only one of these three reached significance when compared with sham stimulation, which was not assessed in Kanai, Muggleton ( 2 ).

A number of factors could potentially explain the discrepant outcomes. One of the main differences in the design of the current study relative to Kanai, Muggleton ( 2 ) was the lack of precise localization of the FEF. To speed application to better suit clinical environments, in the current study we simplified the tDCS protocol by using basic EEG-based measurements to position the FEF electrode, in accordance with Ro, Cheifet ( 24 ) and Ro, Farne ( 25 ). However, there were several other design differences that may have influenced the results. For example, the saccade paradigm used in the current study (adapted from 30) differed from that used in Kanai, Muggleton ( 2 ), in that in their study permanent boxes marked the possible saccade signal locations (where as the saccade signal locations were unmarked in the current study), the fixation dot overlapped with the saccade signal (where as the fixation dot disappeared when the saccade signal appeared in the current study), the fixation duration varied from 300-700 ms (700-1500 ms in the current study), the response period varied from 50-400 ms (50-1500 ms in the current study), the saccade velocity threshold was 28.6°/s (50°/s in the current study), and eye position was sampled at 250 Hz (1100 Hz in the current study). Also, the reference electrode was placed on the shoulder (deltoid muscle) in Kanai, Muggleton ( 2 ) but on the upper arm in the current study. Although these design differences may have influenced the results, none of these design differences should affect performance asymmetrically, and thus they cannot explain asymmetries present in Kanai, Muggleton ( 2 ) but not in the current study. Hence, the use of basic measurements to position the electrodes seems the most likely factor underpinning the discrepant results.

The lack of benefits in older adults came as a particular surprise given that they have far more room for improvement and past research indicates that tDCS can confer greater benefits in older adults ( 19 ). One factor that may have contributed to the failure to induce improvements in saccadic eye movement control in the older adults pertains to age-related increases in cerebral spinal fluid ( 35 ), which can attenuate electric field strength ( 36 ). Another factor that may have reduced the chances of inducing benefits in the older adults is that the tDCS protocol used may not suit older adults due to agerelated changes in brain activation patterns ( 37 ). As reviewed in Dayan, Censor ( 38 ), small electrodes stimulate more focally, which can be beneficial in some circumstances. However, given that older adults normally show widespread prefrontal activation not seen in young adults especially when engaged in higher level cognitive processing (see 13, for a review), focal stimulation may not be optimal to induce pervasive physiological changes necessary to enhance saccadic eye movement control in older adults.

Another factor that may have contributed more generally to the lack of tDCS effects pertains to the spatial distribution of the induced electric field. As demonstrated in Moliadze, Antal ( 39 ), the reference electrode positioning determines the direction of current flow whilst the distance between the electrodes determines where the peak electric field is focused. Given that current passes between the two electrodes, an anode placed over the frontal region and a cathode (i.e., reference electrode) placed over the deltoid muscle or upper arm leads to the current flowing in from the anodal electrode site, passing through the brainstem and the spinal cord, and diffusing at the site of the reference electrode ( 40 ). This tDCS montage, used in the current study and in Kanai, Muggleton ( 2 ), should have resulted in the electric field concentration (i.e., the “hotspot”) being distributed outside of prefrontal regions, roughly around the neck region. Thus, the electrode positions used here and in Kanai, Muggleton ( 2 ) may not be optimal for inducing physiological changes in prefrontal regions.

With respect to developing a tDCS protocol that is more likely to induce physiological changes required to improve functioning, especially in older adults, future studies should take into consideration using a contralateral encephalic reference electrode (e.g., over the forehead or cheek), which should optimize the electric field in prefrontal regions ( 41 ). This arrangement, usually involving a large active electrode over prefrontal cortex combined with a contralateral encephalic reference electrode, has shown promise in a large number of studies that reported improvements in non-oculomotor cognitive functions in older adults ( 42, 43 ). This more typical montage may be worthy of assessment in relation to oculomotor functions as well.

In conclusion, the current study found no evidence that anodal tDCS over frontal subregions improves saccadic eye movement behavior. The failure to produce benefits using a more clinically practical protocol, adapted from Kanai, Muggleton ( 2 ), suggests that localization of the FEF may be necessary for this smallelectrode tDCS protocol to be effective. Future efforts to develop a clinically practical protocol should consider using a larger active electrode and positioning the active and reference electrodes such that the maximally stimulated brain regions are relevant to the functions targeted in the population under study. In addition, a sham stimulation control condition should always be included to enable confirmation that any apparent benefits in active stimulation conditions are attributable to the tDCS.