This study explored the modulatory effects of high-frequency transcranial random noise stimulation (tRNS) on visual sensitivity during a temporal attention task. We measured sensitivity to different onset asynchronies during a temporal order judgment task as a function of active stimulation relative to sham. While completing the task, participants were stimulated bilaterally for 20 min over either the TPJ or the human middle temporal area. We hypothesized that tRNS over the TPJ, which is critical to the temporal attention network, would selectively increase cortical excitability and induce cognitive training-like effects on performance, perhaps more so in the left visual field [Matthews, N., & Welch, L. Left visual field attentional advantage in judging simultaneity and temporal order. Journal of Vision, 15, 1–13, 2015; Romanska, A., Rezlescu, C., Susilo, T., Duchaine, B., & Banissy, M. J. High-frequency transcranial random noise stimulation enhances perception of facial identity. Cerebral Cortex, 25, 4334–4340, 2015]. In Experiment 1, we measured the performance of participants who judged the order of Gabors temporally imbedded in flickering discs, presented with onset asynchronies ranging from −75 msec (left disc first) to +75 msec (right disc first). In Experiment 2, we measured whether each participant's temporal sensitivity increased with stimulation by using temporal offsets that the participant initially perceived as simultaneous. We found that parietal cortex stimulation temporarily increased sensitivity on the temporal order judgment task, especially in the left visual field. Stimulation over human middle temporal area did not alter cortical excitability in a way that affected performance. The effects were cumulative across blocks of trials for tRNS over parietal cortex but dissipated when stimulation ended. We conclude that single-session tRNS can induce temporary improvements in behavioral sensitivity and that this shows promising insight into the relationship between cortical stimulation and neural plasticity.
Humans rely heavily on the ability to temporally discriminate events with high precision and sensitivity. Complex visual, motor, and auditory actions, such as playing sports, driving a car, or playing a musical instrument, rely inherently on processing events as they unfold and acting with the appropriate response.
Thus, humans have developed an innate capacity to detect temporal changes and movement in the visual field; in fact, experimentally, participants detect visual flicker up to ∼50 Hz (Holcombe, 2009; He, Cavanagh, & Intriligator, 1997). However, when asked to report on temporal qualities of the visual field such as the temporal order of stimuli presented with some offset or location of an “oddball” disc flickering out-of-phase relative to distractors temporal resolution slows to about 10 Hz (Agosta et al., 2017; Tyler, Dasgupta, Agosta, Battelli, & Grossman, 2015; Aghdaee & Cavanagh, 2007; Battelli, Cavanagh, & Thornton, 2003). This limited threshold of temporal resolution reflects slow cortical processing of stimuli that demand additional attention resources in conjunction with basic perceptual processes.
Control of temporal attention relies on a large-scale network involving sensory cortex and the canonically defined dorsal and ventral attention networks (for a review, see Corbetta & Shulman, 2011). Importantly, the right temporo-parietal junction (rTPJ) is a critical hub in the “when” attention network (Battelli, Walsh, Pascual-Leone, & Cavanagh, 2008; Battelli, Pascual-Leone, & Cavanagh, 2007). The importance of the right parietal cortex is exemplified in neurological patient literature; damage (typically due to stroke) to this region can result in temporal neglect across the entire visual field (Roberts, Lau, Chechlacz, & Humphreys, 2012; Battelli, Cavanagh, Martini, & Barton, 2003; Battelli et al., 2001). Induced “temporary lesions” using inhibitory TMS over the rTPJ creates similar effects as temporal neglect stroke patients (Agosta et al., 2017; Dambeck et al., 2006; Meister et al., 2006). Moreover, temporary inhibition of the parietal cortex disrupts temporal phase discrimination, resulting in reduced temporal order perception (Woo, Kim, & Lee, 2009).
Recent research has shifted focus to different patterns of neuromodulatory brain stimulation that can potentiate behavioral performance. For example, transcranial direct current stimulation (tDCS) and transcranial alternating current stimulation can boost performance in visuospatial attention tasks (Roy, Sparing, Fink, & Hesse, 2015; Sparing et al., 2009), whereas transcranial random noise stimulation (tRNS) can enhance performance in face processing, numerical cognition, and language acquisition (Pasqualotto, Kobanbay, & Proulx, 2015; Romanska, Rezlescu, Susilo, Duchaine, & Banissy, 2015; Cappelletti et al., 2013). tRNS over the posterior parietal cortex can significantly enhance and prolong behavioral improvements for up to several weeks posttraining (Cappelletti et al., 2013).
Combined with our knowledge of the “when” attention network and the excitatory effects of neurostimulation, we reason that tRNS can potentiate parietal cortical activity and consequently improve behavior on temporal attention tasks. In our experiments, we measured the effect of high-frequency tRNS (hf-tRNS; alternating currents applied in the 100–640 Hz range) over nodes of the “when” attention network while participants were engaged in a temporal order judgments (TOJ) task. We stimulated participants bilaterally over either the TPJ or the human middle temporal area (hMT+). Previous studies demonstrate that TPJ is a cortical hub of temporal processing (Tyler et al., 2015; Battelli et al., 2008). hMT+ is also activated during temporal attention tasks (Tyler et al., 2015; Bueti & Walsh, 2009); however, patients affected by lesions to hMT+ do not present with relative timing deficits (Battelli, Cavanagh, Martini, et al., 2003), and TMS delivered over hMT+ does not disrupt performance in tasks requiring visual timing (e.g., biological motion discrimination; Grossman, Battelli, & Pascual-Leone, 2005). We therefore hypothesized that tRNS over the TPJ will selectively increase cortical excitability in the “when” network and induce cognitive training-like improvements on performance (Romanska et al., 2015). Stimulation over hMT+, however, should not change cortical excitability in a way that affects behavioral performance. We found that timing sensitivity on the TOJ task increased during one session of hf-tRNS over bilateral parietal cortex, but not over hMT+. This rapid effect strongly encourages further research into stimulation paradigms that are not only behaviorally beneficial but also long lasting, thus creating a potential translational application in the clinical population as an aid in neurorehabilitation.
Forty-four individuals with normal or corrected-to-normal vision participated in this experiment (eight men, ages 18–35). Participants were assigned to one of four stimulation conditions: (1) behavioral (n = 11), (2) hMT+ (n = 11), (3) parietal (n = 11), (4) sham (n = 11). Each participant was randomly assigned to a condition group, with the exception that any participant who did not meet standard inclusion criteria based on a brain stimulation questionnaire was automatically assigned to the behavioral group. All participants gave informed, written consent approved by the ethics committee of the University of Trento and received monetary compensation for their participation in the experiment.
Apparatus and Stimulus
The stimulus sequence consisted of a pair of discs (each 1.6° of visual angle) positioned 4° to the left and right of a central fixation cross (Figure 1A). The discs appeared at the start and end of each trial as a stationary texture pattern for 500 msec and then dynamically flickered at a fixed 7.5 Hz between high-contrast black and white in the intervening 1000 msec. This flicker rate was chosen because it is well within participants' temporal discrimination threshold reported in earlier research (Battelli et al., 2007; Verstraten, Cavanagh, & Labianca, 2000). Texture patterns at the onset and offset of the trial masked any cue as to the starting phase of the flickering discs and prevented any resulting perceptual afterimages from the final frame.
The flickering target discs were presented with temporal offset asynchronies of ±75, 66.7, 58.3, 50, 41.7, 33.3, 25, 16.7, 8.3, or 0 msec. During each trial and at least one cycle before the end of the trial, the discs changed into a Gabor for one flicker cycle. The earliest a disc could “change” into a Gabor was 133 msec post-flicker onset. A negative offset indicated that the right Gabor appeared first, whereas a positive offset indicated that the left Gabor appeared first. The Gabors were vertically oriented with a spatial frequency of 3 cycle/deg and a reduced 50% contrast relative to the rest of the flicker cycle. The discs were visible against a uniform gray background (47.5 cd/m2), with a fixation marker positioned in the center of the screen for the duration of the experiment.
All stimuli were displayed on a 22-in. LCD monitor with a 120-Hz refresh rate controlled by a DELL computer equipped with Matlab r2010a (The MathWorks, Natick, MA) and Psychtoolbox 3.0.8 (Brainard, 1997; Pelli, 1997). Participants were seated 57 cm away from the screen in a dark and quiet room and used a chin rest to ensure consistent positioning.
Participants performed a two-alternative forced-choice visual TOJ task. A trial began with a black fixation cross on an otherwise blank screen, and then two discs appeared to the right and left of the cross. Participants were instructed to maintain fixation on the central cross at all times. Each trial lasted 2 sec (500 msec mask at the beginning/end, 1000 msec flicker). During this flicker presentation, the discs change to Gabor patches for 133 msec. At the end of the second stationary texture mask, the discs disappeared (with only the fixation cross remaining on the screen). At this time, participants were prompted to report on whether the left or the right Gabor patch appeared first using either the F key (left first) or the J key (right first). After the participant responded, their response accuracy was displayed on screen for 500 msec (“Sbagliato” [incorrect] or “Giusto” [correct]). The next trial began when the participant pressed the space bar.
Participants first completed one practice block (110 trials; without stimulation) to familiarize themselves with the experiment and to get a baseline performance measurement. After the practice block, those participants in a stimulation group were measured and fitted with electrodes. For the hMT+ and Parietal conditions, 20-min online stimulation began with the start of Block 2 and lasted until the end of Block 5 (Figure 1C). Participants completed four blocks of the experiment with stimulation and one last block without stimulation (all identical to the practice block) for a total of six blocks, 110 trials each. Temporal offset values were pseudorandomly assigned and balanced across the six blocks. Feedback was provided during the course of the entire experiment.
Before testing, all participants were provided with a short introduction to brain stimulation and safety information. After each participant was briefed, they completed a stimulation testing questionnaire and signed the informed consent. At this time, any participant deemed ineligible for stimulation was automatically assigned to the behavioral group. Participants that met all required criteria for stimulation were randomly assigned to one of four stimulation conditions: behavioral only, stimulation over middle temporal cortex (hMT+), stimulation over parietal cortex, and sham stimulation over middle temporal cortex. Stimulation locations were chosen based on previous research that finds unique functional connections between motion sensitive visual cortex (hMT+) and the parietal cortex (the rTPJ) during temporal attention tasks (Tyler et al., 2015).
Stimulation was induced with a DC-Stimulator (Eldith-Plus, NeuroConn). For the hf-tRNS conditions, 1 mA current was applied for 20 min with random alternating frequency deliverance between 101 and 640 Hz. Stimulation was delivered with a fade in/out period of 20 sec at the beginning and at the end of the session. For sham, the electrodes were placed over the hMT+ stimulation site, but the machine was turned off after the fade in/out period. Stimulation sites were identified using the International electroencephalographic 10/20 system for scalp electrode localization (Figure 1B). The center of the saline-soaked electrode was placed bilaterally over PO7/PO8 (left and right, respectively) for hMT+ and sham conditions and bilaterally over P3/P4 for the parietal condition. Electrodes were held in place using a rubber swimming cap.
Participants reported no noticeable sensation resulting from tRNS. Historically, participants cannot tell the difference between active and sham tRNS (Ambrus, Paulus, & Antal, 2010).
To assess sensitivity to temporal order across time, for each block we measured individual participants' precision in guessing the order of the Gabors on the TOJ task at each offset. We then plotted performance at each temporal offset proportionally as “responded that LEFT Gabor appeared first.” Therefore, when the right Gabor was presented 75 msec before the left, this value will be close to 0% (never reported “LEFT” first). As the temporal offset between Gabors approached 0 msec (synchronous presentation), participants' performance worsens as they are no longer able to differentiate the order of onset of the Gabors and reported “LEFT” first trials hover around 50%.
Performance across each individual block was then fitted with a cumulative normal psychometric function (Palamedes Toolbox 1.8.2; Prins & Kingdom, 2009). The resulting slope of the psychometric function shows the relationship between participant sensitivity and the displayed temporal order asynchrony; a steeper slope indicates higher sensitivity (participants are better able to discriminate temporal offsets closer to 0 msec).
Each participant's performance was thus represented as six slope values (one per block). To measure changes in temporal acuity over time, we compared differences in slope over each successive block compared with the initial performance from Block 1, separated by stimulation condition. We used repeated-measures ANOVA (alpha level = .05; statistical analyses were performed using SPSS version 24, IBM, Armonk, NY) to measure changes in slope values over time with block as a within-subject factor and stimulation condition (tRNS) as between-subject factor. Post hoc pairwise comparisons reveal any simple effects between pairs of stimulation conditions not otherwise inferred from the main effects analysis.
Effect of Stimulation on TOJ Performance
Repeated-measures ANOVA revealed no main effect of Stimulation site on TOJ performance, F(3, 40) = 2.467, p = .075, η2 = .157. However, post hoc pairwise comparisons show that parietal stimulation alone caused significantly better performance on the TOJ task than the behavioral (p = .05) and sham (p = .014), but not hMT+ (p = .079; Figure 2). There was no difference in performance among behavioral, hMT+, and sham conditions (which explains the lack of a main stimulation effect). This result suggests that, only under active parietal stimulation, participants display heightened temporal acuity at increasing shorter TOJ offsets.
Effect of Stimulation over Time
Based on previous reports that the effects of tRNS are not immediate (typically, it can take approximately 15–20 min into stimulation to see significant effects; Antal, Ambrus, & Chaieb, 2014), we analyzed performance over time to see when stimulation started showing behavioral effects. Repeated-measures ANOVA did not find a significant difference across blocks when collapsed across stimulation conditions, F(5, 200) = 2.140, p = .062, η2 = .051; however, simple effects analysis shows that stimulation over parietal cortices significantly affected performance across blocks, F(5, 36) = 2.495, p = .049, η2 = .257; performance on Block 1 was worse than Blocks 2, 3, 4, and 5 (p = .02, .044, .008, .004, respectively; Figure 2). This shows that performance on the TOJ task steadily increased over time for parietal stimulation only, a trend that was not present in behavioral, F(5, 36) = .690, p = .634, η2 = .087, hMT+, F(5, 36) = 1.168, p = .344, η2 = .140, or sham, F(5, 36) = .899, p = .492, η2 = .111, conditions. This effect of parietal stimulation was gradual, with peak performance measured approximately 15 min into stimulation. Blocks 4 and 5 display higher mean performance than Block 6 for parietal stimulation (p = .02 and p = .048, respectively). As shown in Figure 2, this effect is due to a rapid decline in performance between Blocks 5 and 6. This effect may be indicative of a buildup of cortical excitability dependent on consistent stimulation, an effect that disappeared when stimulation ended after Block 5 (Pirulli, Fertonani, & Miniussi, 2013).
Experiment 1 demonstrates that single-session tRNS over bilateral parietal cortex can improve overall acuity on a TOJ task. Moreover, the data show that the effect is very rapid and can evolve within one session only, a clear indication of cortical plasticity in the adult cortex. However, because the effects of tRNS lasted only during online stimulation, single-session tRNS might not be enough to elicit any long-term plastic changes.
Experiment 1 demonstrated a general increased sensitivity in perceived onset asynchrony (SOA) across a range of offsets. Experiment 2 was developed to specifically target whether participants effectively become better at temporally separating the onsets of discs they originally perceive to be simultaneous (point of subjective simultaneity). We also examine whether stimulation can alter the temporal range in which participants can accurately discriminate temporal order tasks across left and right visual fields based on the widely reported left visual hemifield bias found in spatial and temporal attention literature (Matthews & Welch, 2015; Śmigasiewicz, Asanowicz, Westphal, & Verleger, 2015; Bosworth, Petrich, & Dobkins, 2012; Śmigasiewicz et al., 2010). A smaller, more sensitive temporal range (or “window”) indicates that participants identify the targets as simultaneous only within a very small temporal offset range versus a larger temporal window in which targets are still identified as simultaneous even with a wider range of offset values.
Twenty-eight individuals participated in Experiment 2 (three men, ages 18–30). Participants were assigned to one of four stimulation conditions: (1) behavioral (n = 7), (2) hMT+ (n = 7), (3) parietal (n = 7), and (4) sham (n = 7). Exclusion criteria were identical to that in Experiment 1. All participants gave informed, written consent approved by the ethics committees of the University of Trento and received monetary compensation for their participation in the experiment.
Apparatus and Stimulus
The stimulus presentation for Experiment 2 was identical to that of Experiment 1, with the exception that the Gabors were presented at one onset asynchrony only, instead of the full range in Experiment 1 (Figure 1A).
The offset at which the Gabors were presented was determined by measuring each individual participant's 50% threshold for offset detection (the point at which participants perceive the discs as simultaneous and thus guess which disc appeared first). Participants completed the same temporal order task as Experiment 1, with the exception that the offset value changed trial to trial, depending on the accuracy of the previous trial. This value was determined using a 1–1 double interleaved staircase procedure, in which a single incorrect response decreased task difficulty (increases the offset gap between left and right Gabor presentations) whereas a single correct response increased task difficulty (decreased offset gap). The initial offset asynchrony of the discs varied randomly between 9 and 42 msec, with step sizes of 8.3 msec. The staircases terminated after a combined 15 reversals of the staircase parameter, with an average of 68.1 total trials across our participants (SD = 9.73), and 50% threshold of 30.96 msec (SD = 12.04).
After completion of the consent form and brain stimulation questionnaire, all participants completed the threshold discrimination block (Block 1; without stimulation) to familiarize themselves with the experiment and measure the 50% threshold estimate for onset asynchrony discrimination. The rest of Experiment 2 proceeded identically as Experiment 1, with stimulation electrodes being placed on participants' scalps between the threshold block and Block 2 (Figure 1B and C).
Participants completed five blocks of the experiment using their individual threshold for a total of six blocks with 110 trials each. Presentation of Gabor order (left first or right first) was pseudorandomly determined and balanced across blocks (Figure 1C). Feedback was provided during the course of the entire experiment.
Stimulation procedure for Experiment 2 was identical to Experiment 1 (Figure 1B).
To track changed temporal sensitivity over time, we measured changes in performance for each successive block relative to the initial practice block. For the hemifield analysis, we divided trials in Experiment 2 into two groups and analyzed the effect: Left Target First, in which the left Gabor preceded the right, and Right Target First, in which the right Gabor preceded the left. To quantify the effect of stimulation time and hemifield effect on shifting point of subjective simultaneity, we ran a repeated-measures ANOVA (alpha = .05) with the within-subject factors being Block and Hemifield and the between-subject factor as Stimulation condition. We separately measured performance in each block for trials with the left and right Gabor presented first and then used the change in accuracy over time as our dependent variable. An increase in accuracy over time indicates that participants have shrunk their “temporal synchrony” window and therefore can correctly discriminate the targets at a small offset. Therefore, this measure is presented as percent difference in accuracy. Statistical analysis was performed using SPSS (version 24, IBM).
There was no effect of stimulation on overall performance, F(3, 24) = .159, p = .923, η2 = .020, for Experiment 2. When we combined all trials across the full visual field, temporal sensitivity did not improve over time; average accuracy rate at the end of the experiment only varied, at maximum, 4% from the beginning of the experiment. This held true across all stimulation conditions (behavioral p =.905; hMT+ p = .956; parietal p = .843; sham p = .949; Figure 3A–D).
However, there was a significant difference in performance between the left and right visual field, collapsed across blocks and stimulation conditions, F(1, 24) = 5.06, p = .034, η2 = .174, with participants performing much better on trials when the left Gabor was presented before the right. This difference across visual field is only reflected in the parietal stimulation condition (Figure 3C; p = .002). When Gabors appeared in the right visual field first, participants on average performed 10.5% worse than performance averaged across visual fields; when Gabors appeared in the left visual field (LVF) first, participants made approximately 5.5% more accurate judgments than average (Figure 3E). The stimulation effect was quicker than Experiment 1, with acuity reaching peak between Blocks 2 and 3 (approximately 5–10 min poststimulation onset). Performance did not gradually increase as stimulation time increased, but rather remained at a steady rate after Block 3.
Overall, these results show that, although average performance across visual fields on the TOJ task did not improve, parietal stimulation increased sensitivity in the left visual field, such that participants could better discriminate temporal offsets previously interpreted as synchronous (decreasing their temporal synchrony “window”). This improvement dovetails with previous research that shows left visual field bias during temporal attention tasks (Matthews & Welch, 2015).
The goal of our study was to measure whether one 20-min hf-tRNS session improves the resolution at which participants could discriminate the sequence of events. In Experiment 1, we found that hf-tRNS over bilateral parietal cortex increased behavioral precision on a TOJ task. Participants were better able to distinguish smaller temporal offsets with parietal stimulation relative to our control conditions (no stimulation, sham stimulation, and middle temporal stimulation). Hence, visual temporal acuity significantly improved in one 20-min hf-tRNS session.
In Experiment 2, parietal tRNS increased temporal acuity in the left visual field when participants were tested at their individual subjective synchrony threshold. As time (and stimulation) progressed, participants became increasingly better at discriminating onsets when the left Gabor was presented before the right. Conversely, performance slightly worsened when the right Gabor was presented before the left. This suggests that parietal stimulation differentially affects sensitivity across visual fields, a phenomenon not uncommon in attention literature (Chambers, Payne, Stokes, & Mattingley, 2004; Hilgetag, Théoret, & Pascual-Leone, 2001).
Combined, these results show that tRNS over parietal cortex affects the processing of visual information, potentially through neuroplastic-like reinforcement of connections to and from the TPJ. Coupled with neurorehabilitation efforts, tRNS may prove a useful tool in facilitating improvement in attentional control in patients with stroke-induced attentional neglect.
We reason that direct current (DC) stimulation effects are selective rather than widespread, propagating out from the site of stimulation in a specific network pattern. The parietal stimulation sites we chose were based off previous functional (Claeys, Lindsey, De Schutter, & Orban, 2003) and neuropsychological work (Agosta et al., 2017; Battelli et al., 2007) that have unpacked the cortical “when” network, the central hub of which is TPJ, inasmuch that trauma to this region impacts connected cortical areas (Battelli et al., 2008). Because of the nature of our stimulation technique, we cannot definitively claim that tRNS specifically targets the TPJ; indeed, it is very possible that larger parts of the temporal and parietal cortices were affected. However, because of the extensive literature that supports the involvement of the TPJ specifically in temporal attention and the fact that our task is a temporal order task, we conclude that this was a major area affected by stimulation.
Indeed, we find that a small difference in area of electrode application (between P3/4 and PO7/8) was enough to induce very different behavioral results. Whereas stimulation over parietal cortex boosts performance, stimulation over hMT+ had no significant effects. Both the rTPJ and bilateral hMT+ are known to be active during temporal discrimination tasks (Tyler et al., 2015), with enhanced connectivity between these two regions when participants are focused on the task. Therefore, we suggest that connection between these regions were controlled in a top–down pattern, with the TPJ influencing the motion-sensitive hMT+ and boosting perceptual sensitivity toward stimuli quickly alternating in time.
The TPJ is linked to a variety of attention-related maintenance tasks, including distractor suppression (Painter, Dux, & Mattingley, 2015), attention reorienting (Chang et al., 2013), and vigilance (Malhotra, Coulthard, & Husain, 2009). tRNS over this region can also act as an exogenous boost to attentional arousal, allowing participants to better focus on a continuous selective attention task (Mauri, Miniussi, Balconi, & Brignani, 2015). Indeed, periods of increased attentional arousal can elicit fast RTs on a simple discrimination task (Bagherli, Vaez-Mousavi, & Mokhtari, 2011) and enhance visual perception (Zeelenberg & Bocanegra, 2010). Anodal tDCS over the rTPJ can facilitate faster visual search and target detection (Clark et al., 2012). In line with our results, application of tRNS over parietal cortex likely induces a heightened state of cognitive arousal within this network, giving participants overall better visual sensitivity to the temporal order of events.
Mechanisms of tRNS
Our results dovetail with previous research that finds that tRNS can be a task-specific modulator, directly enhancing cortical hubs in proximity at the site of stimulation and likely propagating to synchronous networks. Terney and coworkers were the first to show improvement in an implicit motor learning task after M1 stimulation (Terney, Chaieb, Moliadze, Antal, & Paulus, 2008). Subsequent studies found that tRNS over V1 modulates performance on an orientation discrimination task (Fertonani, Pirulli, & Miniussi, 2011), enhances face perception when delivered over lateral occipital complex (Romanska et al., 2015), and encourages acquisition of a foreign language and numerosity discrimination when administered over the parietal cortex (Pasqualotto et al., 2015; Cappelletti et al., 2013), and over dorsolateral pFC increases speed in arithmetic learning (Snowball et al., 2013).
There are a few theories on the mechanisms by which tRNS modulates cortical excitability. One prevalent hypothesis suggests that tRNS operates through a continuous cycle of potentiation of sodium-controlled ion channels (Chaieb, Antal, & Paulus, 2015; Terney et al., 2008) to create heightened cortical excitability. In turn, this encourages long-term potentiation-like neuronal changes that enhance connections between the area of stimulation and connected hubs (Snowball et al., 2013; Fertonani et al., 2011). In our task, this could create a stronger barrier against irrelevant flicker information until the targets appear (similar to distractor suppression in a rapid serial visual presentation [RSVP] task). There is evidence that deactivation of the rTPJ leads to better target detection, only coming online in response to relevant behavioral stimuli and thus “filtering” out irrelevant information (Corbetta, Patel, & Shulman, 2008; Shulman, Astafiev, McAvoy, d'Avossa, & Corbetta, 2007). With stimulation, we could be tapping into this deactivation pattern, causing less overall “noise” to be introduced into the network and create higher target sensitivity results.
Another theory suggests that tRNS injects a low amount of noise into the system, counterintuitively increasing information in a nonlinear fashion and thus pushing weakened signals above “detection” thresholds (in our case, small temporal asynchronies; Stocks, 2000). This theory of stochastic resonance leads to higher synchrony of neuroscillations, thus creating strong links between concurrently firing neurons (van der Groen & Wenderoth, 2016; Wiesenfeld & Moss, 1995). For our study, this implies that stimulation increased the temporal neural synchrony between rTPJ and the connected “when” network, thus increasing sensitivity toward temporal offsets that previously may not have reached attentional awareness. Previous stimulation research has suggested that low-intensity TMS can induce stochastic resonance-like behavioral facilitation (Schwarzkopf, Silvanto, & Rees, 2011). We can therefore hypothesize that, without the additional noise, detection threshold for temporal order remains balanced across visual fields; there is a “weak” temporal signal from the task when the Gabor offset is small. With the added stimulation, it seems that only information from the left visual field is boosted (as seen from our Experiment 2 results). As for the right visual field, either added noise fails to boost signals above threshold or actually oversaturates the system with information, effectively “drowning out” the signal (see Moss, Ward, & Sannita, 2004, for a review).
Whichever theory holds true, signals from previously undetected target onset asynchronies are boosted under the influence of tRNS, which leads to more sensitive discrimination of TOJs. We can conclude that bilateral tRNS over the parietal cortex enhances attentional focus toward temporal attention tasks by gating the flow of distracting temporal information through the motion-sensitive “when” network. However, future studies do need to uncover the nature of the rTPJ neural activity, as some studies report that deactivation of this area increases attention vigilance (Wu et al., 2015; Shulman et al., 2007) whereas others report the opposite (Kincade, Abrams, Astafiev, Shulman, & Corbetta, 2005).
In Experiment 2, we found that, with bilateral stimulation over parietal cortex, participants were more accurate in judging temporal differences when the left disc was presented before the right. In attention literature, there is a phenomenon known as the “left visual field advantage,” where participants are better at attention tasks in that hemifield (Bosworth et al., 2012). Specifically within temporal attention, during dual RSVP tasks, the second target in an attentional blink paradigm was better identified in the left visual field (Śmigasiewicz et al., 2010; Verleger et al., 2009; Holländer, Corballis, & Hamm, 2005). Similarly, during simultaneity judgments (Matthews, Vawter, & Kelly, 2012), participants' miss rates are lower for the left visual field than the right.
This effect is thought to be due to hemispheric competition for attentional control between the right and left parietal cortices. In typical participants, the right parietal cortex exerts control over the left, creating a slight bias toward the left visual field (Szczepanski, Konen, & Kastner, 2010); however, in patients with hemispatial neglect, right parietal damage creates an imbalance in interhemispheric competition, in which inhibition of the right parietal cortex creates hyperactivation of the left parietal cortex and shifts visual field biases toward the right (de Haan, Karnath, & Driver, 2012; Corbetta & Shulman, 2011). Similarly, TMS and bilateral tDCS over the right parietal cortex mimics this deficit and creates disruptions to interhemispheric balances, creating a rightward shift in attentional awareness (Filmer, Dux, & Mattingley, 2015; Petitet, Noonan, Bridge, O'Reilly, & O'Shea, 2015; Giglia et al., 2011; Battelli, Alvarez, Carlson, & Pascual-Leone, 2009; Dambeck et al., 2006). Unilateral DC stimulation studies have replicated the LVF effect, albeit with mixed results. In Sparing et al. (2009), anodal tDCS over right posterior parietal cortex facilitated LVF visual search while cathodal tDCS inhibited, and Bolognini, Olgiati, Rossetti, and Maravita (2010) found decreased RTs in the LVF for a target detection task using anodal tDCS. However, in a more recent study, anodal tDCS was also found to inhibit LVF sensitivity when targets were presented bilaterally (Filmer et al., 2015).
Combined, this research suggests that the interhemispheric bias toward the LVF can be exacerbated under excitatory tRNS. To our knowledge, ours is the first bilateral DC stimulation study that increases the left visual field advantage in normal participants while simultaneously inhibiting the right visual field. This phenomenon may indicate that tRNS more strongly modulates the rTPJ; hence, the larger right hemisphere-biased “when” attention network.
Overall, we find that bilateral tRNS over the parietal cortex boosts temporal attention across the visual field, with a larger advantage emerging in the LVF. This excitability translates into more sensitive perceptual acuity for onsets of temporal targets. This research provides important insight into how we may use stimulation to improve cognitive behaviors. Neurostimulation techniques are becoming an increasingly relevant research field for neurorehabilitation, with early research showing promising results (Agosta, Herpich, Miceli, Ferraro, & Battelli, 2014; Sparing et al., 2009). Through increased and prolonged states of cognitive arousal, neuroplastic-like reinforcement of the weakened or damaged cortical connections caused by focal damage to the parietal cortex will hopefully relieve some of the behavioral and cognitive impairments that these patients suffer through. By increasing cortical excitability in the attention-controlling parietal cortex, patients may be able to ameliorate the shift in interhemispheric balance caused by trauma.
We thank J. A. Assad for comments and discussions on an earlier version of the manuscript. The research was funded by the Autonomous Province of Trento. Call “Grandi Progetti 2012,” project “Characterizing and improving brain mechanisms of attention – ATTEND.”
Reprint requests should be sent to Sarah C. Tyler, Department of Psychology, University of California, San Diego, 9500 Gilman Drive #0109, La Jolla, CA 92093-0109, or via e-mail: email@example.com.