The current study provides evidence of the neuromodulatory effects of tDCS to the right inferior frontal gyrus on the P3 component and underlying oscillatory activity during a waiting impulsivity task (CPRT). A decrease of target and cue-P3 amplitude elicited in CPRT during tDCS over rIFG was found. Moreover, the reduction in target-P3 amplitude during active stimulation was combined with a simultaneous reduction in delta activity for the same time-window. Regarding behavioral analysis, there was a significantly higher k in the small amounts condition after active tDCS in comparison with sham. However, no modulatory effects of tDCS over rIFG in waiting impulsivity measures from CPRT were found.
Electrophysiological correlates
In the present study, anodal tDCS to the rIFG decreased the target-P3 amplitude and underlying oscillatory activity (namely delta activity) during waiting impulsivity response. However, there is a growing body of evidence suggesting that anodal tDCS in frontal areas increases P3 amplitude during performance of cognitive paradigms involving attentional and working memory processes 20, and, a decrease in P3 amplitude during inhibitory control paradigms 22,23. Thus, the differential effects of tDCS on P3 is related to the functional role of each cognitive task and its underlying neuronal substrates 20.
Furthermore, these findings underpin the relationship between P3 and the delta/theta activity at the same time-window observed in several cognitive tasks 18,36 (Inter-Trial Phase Coherence (ITPC) analysis and additional discussion about cue-P3 in Supplementary Materials). In fact, a recent study showed an enhancement of P3 amplitude during a visual oddball paradigm after the entrainment of delta/theta frequency bands through the application of transcranial Alternating Current Stimulation (tACS) 37. However, contrarily to our hypothesis, theta activity was not modulated concurrently to both cue and target-P3 amplitude. On the other hand, tDCS modulated delta activity concurrently with the target-P3 amplitude, suggesting a potential inter-dependency between both electrophysiological markers and its importance during impulsive behaviors. For instance, a study decreased impulsive eating behavior in rats through a closed-loop system that triggered a responsive neurostimulation in the nucleus accumbens every time delta activity was excessively increased during reward anticipation 38. Likewise, delta activity and P3 amplitude in parietal region are also enhanced during the anticipation of rewards when compared to neutral trials 19. This is of particular interest because it was showed before a positive correlation between the cue-P3 amplitude and the activity in the ventral striatum (including nucleus accumbens), which is a core structure for reward processing and impulsive choice 11. Similarly, neuronal activity observed left ventral striatum activity during an inhibitory control task was negatively correlated with the rIFG 39. Thus, this might suggest that tDCS over rIFG might not only result in the reduction of P3 amplitude and delta activity, but also in the decrease of ventral striatum activity. This is in accordance with the behavioral results as well, given that tDCS over rIFG increased the choice for immediate rewards, which, for its hand, it has been associated with reduced activity on ventral striatum 40.
Furthermore, the anticipatory (i.e., cue) and consummatory (i.e., target) P3 in paradigms with monetary incentives are strongly related with reward processing, although they show different patterns. The target-P3 amplitude is greater when preceded by cues that predict win and loss of monetary compensation, which suggests its involvement during reward and punishment processing 12. On the other hand, mixed effects in relation to the cue-P3, such as an enhancement after reward cues 11–13, loss cues15, or both 14 has been shown in literature. Similarly, as previously mentioned, cues that predict rewards elicited an enhancement of delta activity in parietal areas (and cue-P3 amplitude) when compared to neutral cues 19. Although, most of previous studies did not show a significant relation between cue-delta activity and delay discounting, a recent study demonstrated that the increase of evoked delta after stimulus during delay discounting paradigm was associated with the choice of delayed and larger rewards 41. Therefore, the decrease of P3 amplitude and delta activity might indicate a modulation in the impulsive choice identified in our delay discounting results (but not found in CPRT and SSRTT).
Behavioral Outcomes
tDCS over rIFG did not impact the CPRT outcomes, namely, the number of premature responses, release time, and total earned money. These findings suggest that, although previous studies suggested the involvement of the rIFG in inhibitory control abilities 22,42, it might not be critically involved in waiting impulsivity 8. Specifically, we expected an increase in tonic inhibitory process, thus, less premature responses. Our results did not support this hypothesis. Nonetheless, several reasons might be pointed out to explain the lack of tDCS effects in waiting impulsivity. First, tDCS over rIFG might show greater effect in reactive inhibition than tonic inhibitory response involved in premature responses. Indeed, a recent meta-analysis exploring the effects of tDCS in both inhibitory processes showed a significantly larger effect size in reactive (e.g., SSRTT) than tonic inhibition (e.g., GNG task) 4. Therefore, a smaller effect of tDCS over rIFG should be expected in premature response given their association with proactive stopping 8. Additionally, the rIFG neural circuits involved in both reactive and proactive inhibition follow different pathways. An indirect pathway has been related to proactive inhibition, which connects the rIFG with globus pallidus through the dorsal striatum, whilst reactive inhibition is related with a hyperdirect pathway from the rIFG and pre-SMA to STN by-passing the striatum 43. Consistent with this notion, the increase of premature responses was associated with lower connectivity within structures relevant for motor inhibition, such as the STN and ventral striatum 9. Therefore, differences in neural pathways might influence how tDCS affects the rIFG based on the network-dependent activity related to the CPRT 44. This is supported by the absence of transfer effects from the waiting impulsivity task to the motor inhibition performance evaluated in the SSRTT. Nonetheless, this hypothesis is not in line with previous literature, given that several studies targeting the same area showed an enhancement of proactive inhibitory processes 22,45–48.
Another explanation is that tDCS might increase the proactive inhibition, but without any consequence to premature responding, given that this dissociation was already observed in literature 9,49. Waiting and stopping has been suggested to be different constructs within the impulsivity realm 6, given that they rely on different cortico-striatal connections between the DLPFC and ventral striatum for waiting processes and between the IFG and dorsal striatum for stopping 7,50. Furthermore, the differences of reward/punishment system might undermine any conclusion about the tDCS effect on premature responses or other outcomes from CPRT as suggested in Pearson correlation. Specifically, when it was harder to gain money, participants incurred in more premature responses. The baseline block was performed in the beginning of each section and subsequently the system of reward/punishment was updated in each session (see Limitations).
Moreover, the preference for immediate and smaller rewards observed in the MCQ-27 might be explained by the activation of concurrent neuronal circuitries between waiting impulsive action and delay discounting 6. This is of particular interest given that both processes depend on ventral striatum even though they share different pathways. Specifically, waiting impulsivity relies on the connectivity of STN with ventral striatum and subgenual cingulate cortex 9, whereas increased magnitudes of delayed rewards were associated with activation of mesolimbic pathways through the ventral striatum, medial prefrontal cortex, and posterior cingulate cortex 40. In line with this, studies have shown an effect of tDCS over the DLPFC in the dopamine release in the ventral striatum 51,52, which might explain the transfer effect of tDCS to the delay discounting assessment. Similarly, a neuroimaging study showed that the activity observed in the rIFG was negatively correlated with the activity found in the left ventral striatum 39. Therefore, the application of anodal stimulation over rIFG might result in lower activation in ventral striatum and consequently the increase of the k observed in the MCQ-27. Nonetheless, to the best of our knowledge, this study was the first to test the effect of tDCS over rIFG in delay discounting.
In general, the tDCS transfer effects were only observed in impulsive choice (i.e., delay discounting) that partially shares neuronal circuits with waiting impulsive action 6. Therefore, the modulation of the neuronal circuits related with the waiting impulsivity is in line with the tDCS model of the network activity–dependent model 44. On the other hand, the lack of transfer effect in the inhibitory control task (i.e., SSRTT) might suggest the dissociation between waiting and stopping impulsivity 53 or between impulsive choice and action 54.