Exp Brain Res DOI 10.1007/s00221-016-4560-5

RESEARCH ARTICLE

The effects of transcranial direct current stimulation over  the dorsolateral prefrontal cortex on cognitive inhibition Shlomit Metzuyanim‑Gorlick1 · Nira Mashal1,2 

Received: 10 December 2015 / Accepted: 12 January 2016 © Springer-Verlag Berlin Heidelberg 2016

Abstract  The present study examines the effects of bilateral transcranial direct current stimulation (tDCS) over the dorsolateral prefrontal cortex (DLPFC) (anodal over left and cathodal over right DLPFC). This study describes the long-term effects of tDCS on cognitive inhibition, using the Hayling task. Twenty volunteers participated in the study and were assigned to either an active or a sham group. Participants heard sentences with the final word missing. They were asked then to complete the sentence with a word that either is appropriate in the context of the sentence (initiation condition) or is completely unrelated in this specific context (suppression condition). All participants performed a baseline Hayling task followed by six stimulation sessions. Subsequent to completion of these stimulations, we assessed immediately Hayling performance and re-assessed this performance 1 month. The results indicate a significant decrease in the number of errors in the active group, but only in the suppression condition that continued for 1 month after the sixth stimulation. The current findings suggest that tDCS can improve cognitive inhibition for the long-term in healthy adults and that the DLPFC has a special role in selecting the correct response and suppressing irrelevant semantic information. Keywords  DLPFC · Hayling task · Cognitive inhibition · tDCS

* Nira Mashal [email protected]; [email protected] 1

The School of Education, Bar-Ilan University, Ramat Gan, Israel

2

Gonda Multidisciplinary Brain Research Center, Bar-Ilan University, Ramat Gan, Israel



Introduction Cognitive control includes a set of functions serving to configure the cognitive system for the performance of specific tasks, especially in challenging and non-routine situations (Botvinick et al. 2004). This set of functions contains specific skills such as attention, working memory, planning, and suppression of actions or thoughts to respond correctly (Carter et al. 1998; Crandall et al. 2015; Cohen et al. 1990; Miyake and Shah 1999; Shallice 1988). A major constituent process of cognitive control involves cognitive inhibition. The term cognitive inhibition refers to the ability to control or suppress irrelevant responses and to adopt relevant responses in place of irrelevant ones. The current study focuses on the capacity to detect and filter out irrelevant semantic information within a stimulus set, with the aim to test the effects of transcranial direct current stimulation (tDCS) on improving cognitive inhibition in healthy participants. Several tasks such as the Flanker task and the Stroop task tap into cognitive inhibition. Blasi et al. (2006) explored the region of the brain involved in a modified version of the Flanker task. The participants were instructed to monitor and suppress a response when a central target arrow was flanked by two arrows pointing in the opposite direction of the central arrow on either side. The brain regions that were activated during monitoring and response inhibition included the dorsal lateral prefrontal cortex (DLPFC), dorsal anterior cingulate cortex (dACC), parietal cortex (PC), and the ventral lateral prefrontal cortex (VLPFC). Findings from brain imaging studies which used the Stroop task also point to the involvement of the DLPFC in response inhibition. For instance, more activation in the DLPFC was found when participants named the color of the word (a task that

13



requires cognitive inhibition) compared with low activation that was seen in this area during word reading (a task that elicits an automatic reaction) (MacDonald et al. 2000). These findings suggest thus that the DLPFC may play a major role in cognitive inhibition (Daskalakis et al. 2008; Kopell and Greenberg 2008; Tekin and Cummings 2002; Yang et al. 2010). In addition to cognitive inhibition, the DLPFC has been implicated in various cognitive control processes. Accumulated evidence from neuroimaging studies supports the role of a brain network including the DLPFC, anterior cingulate cortex, and the parietal cortex (PC) in cognitive control (Botvinick et al. 2001; Carter et al. 1999; Cohen et al. 2000; Yarkoni et al. 2005). In a study that tested brain activation in 344 participants while performing a comprehensive battery of neuropsychological tasks (Wisconsin card sorting test, Trail making test, Stroop test, Lowa gambling task), activation was observed in the DLPFC. This activation probably occurred because this region maintains goals by flexibly adjusting attention and working memory resources to changing environment and task demands (Gläscher et al. 2012). Additional activation was also observed in the rostral ACC when flexible shifting between cognitive tasks and response sets was needed. Overall, these findings corroborate other studies that argue that the DLPFC is activated in the following conditions: controlling response inhibition (Blasi et al. 2006; Garavan et al. 2002; Kelly et al. 2004); maintaining rules that guide reaction (Asaad et al. 2000; Lesh et al. 2011; Watanabe 1990, 1992); and updating and selecting the essential information to complete the task (Bunge et al. 2001; Garavan et al. 2002). Based on these findings, Lesh et al. (2011) presented a model for cognitive control. In this model, the DLPFC receives information from both the PC and the ACC to evaluate and select the appropriate response. In this model, each region has specific functions: The PC is responsible for transmitting information to the DLPFC regarding the need to focus attention and allow access to former knowledge about past experience of the stimuli and expected results (Bunge et al. 2002, 2003; Miller and Cohen 2001; Posner and Petersen 1990). The ACC is responsible for alerting the DLPFC when interference is detected (Egner and Hirsch 2005; Kerns et al. 2005; MacDonald et al. 2000). Thus, based on the information received by the PC and the ACC, the DLPFC guides the selection of the relevant response or the suppression of the inappropriate one. Another task that engages cognitive inhibition is the Hayling task (Burgess and Shallice 1997). In this task, participants are asked to either complete the sentence with a word that fits the sentence (initiation condition) or complete the sentence with a word that is completely unrelated to the sentence (suppression condition). The suppression condition requires cognitive inhibition as several missing words

13

Exp Brain Res

are activated and need to be suppressed to select an unrelated word. Using positron emission tomography (PET), Nathaniel-James and Firth (2002) examined the DLPFC function while performing the Hayling task. They found increased activation in the DLPFC in the suppression condition, in comparison with the initiation condition. Royer et al. (2009) investigated brain activity during inhibitory tasks in 19 patients with schizophrenia and 12 healthy controls using the Hayling and the n-back tasks. Both groups activated the prefrontal cortex, temporal cortex, and parietal cortex while performing the Hayling task. In another study, the effective connectivity between left middle temporal region and left middle frontal region was tested using dynamic causal modeling during Hayling task performance (Allen et al. 2008). Response suppression elicited greater activation in the left prefrontal cortex. Furthermore, greater effective connectivity between the left temporal and prefrontal regions was observed during response suppression. This finding is probably because semantic and lexical information was activated in posterior temporal cortex and transmitted to prefrontal regions. In the prefrontal cortex, this information is manipulated to select the relevant response. Over the years, several studies investigated the effect of tDCS on cognitive control (e.g., Boggio et al. 2010; Fecteau et al. 2007; Jacobson et al. 2011). For example, a recent study showed improvement in verbal insight problem-solving using tDCS (Metuki et al. 2012). The results show increased solution recognition for difficult problems in the active condition, compared with the sham condition. This finding is in accord with the study of Wolkenstein and Plewnia (2012). This study showed enhanced performance in a delay-response working memory task using tDCS over the left DLPFC in both healthy participants and participants with a major depressive disorder. Thus, both studies show that a unilateral montage with anodal stimulation over the left DLPFC improves cognitive control performance. Recent studies also point to long-term effects of consecutive tDCS sessions (Dockery et al. 2009; Ferrucci et al. 2009; Jones et al. 2015). For instance, Jones et al. (2015) investigated the improvement in working memory (WM) in older adults after 10 sessions of tDCS stimulations combined with WM training. WM ability was tested at three time points: before the first stimulation, after the tenth stimulation, and a month after the tenth stimulation. At each time point, participants completed three WM tasks (digit span, Stroop, and n-back). The results showed improvement in WM after 10 sessions in both active and sham groups. However, a month after the tenth session, only the active group showed higher performance across trained and transfer tasks. The long-term effects of a series of tDCS sessions were also established in studies with non-healthy participants: post-stroke patients or patients

Exp Brain Res

with psychiatric disorders (de Aguiar et al. 2015; Ferrucci et al. 2009; Lifshitz Ben Basat et al. 2016; Marangolo et al. 2013; You et al. 2011). Previous studies suggested that anodal tDCS increases the N-methyl-d-aspartate (NMDA) receptor activity (Nitsche et al. 2004) and reduces local concentrations of the inhibitory neurotransmitter gammaaminobutyric acid (GABA). This reduction, in turn, increased synaptic plasticity and perhaps produced longterm effects (e.g., Rroji et al. 2015). The present study investigates the effect of tDCS over the left DLPFC on cognitive inhibition using the Hayling task. Furthermore, we tested the long-term effects of six sessions of brain stimulations, 1 month after the last stimulation session. In accordance with neuroimaging studies pointing to the involvement of the left DLPFC in cognitive control tasks, we expect that the stimulation would improve task performance. Specifically, we expect a reduced number of errors and faster reaction times in the suppression condition in the active group, in comparison with the sham group. In accordance with previous tDCS studies pointing to long-term effects of consecutive brain stimulations, we hypothesize that the improvement will last in the active group only, for 1 month after the last stimulation.

Methods Participants Twenty right-handed volunteers (11 females), ranging in age from 21 to 41 years (mean = 30.8, SD = 6.20) and native Hebrew speakers, participated in the study. Twelve (eight females) participated in the active stimulation condition and eight (three females) in the control group (sham). No differences in gender were found χ2 (1) = 1.65, p = .20. All participants had 12 years of education. The two groups were matched according to both the semantic and the phonemic fluency tests (Kavé et al. 2010). In the phonemic fluency test, the participants were permitted 60 s to name as many words as possible that begin with a specific letter (b, g, sh). In the semantic fluency test, the participants were asked to name as many words as possible which belong to a specific category (animals, vehicles, fruits, and vegetables). No significant differences between the study and the

Table 1  Summary of mean age and number of words (SD) in the fluency tests

control group were found in the performance of the fluency tests. These results are summarized in Table 1. The recruitment of participants and research protocols were approved by the Ethics Committee in the School of Education, Bar-Ilan University. All participants signed a consent form that explained the study aims and procedures. The Hayling task The Hayling task (Burgess and Shallice 1997) included two conditions: response initiation and response suppression. In the initiation and in the suppression conditions, participants heard a sentence with the final word omitted. In the initiation condition, participants had to generate a word that is compatible with the context of the sentence. For example, the word that fits the sentence I went to the doctor because I was… is “sick.” In the suppression condition, participants were asked to complete the sentence with a word that is incompatible and completely unrelated in the context of that sentence. For instance, I brush my teeth every morning with… is “wheel.” Stimuli construction Since the original stimuli were in English, we created a new stimuli set in Hebrew (Burgess and Shallice 1997). Accordingly, a pool of 200 sentences with the final word missing was created by the authors. A pretest study was conducted to select the final word of each sentence. Forty judges (20–40 years old) were presented with the 200 sentences in which the final word was omitted. They were asked to complete the sentence with one appropriate word in the context of the sentences. The 60 sentences which the same final word was selected by at least 90 % of the judges were selected for the study. The 60 sentences were divided into four lists of 15 sentences each that were used for the initiation and the suppression condition. Task procedure The participants listened to a female voice reading a sentence through headphones and had 3000 ms to complete the missing word. Words completed after this time limit were considered misses. For each participant in each condition, we counted the number of correct responses, misses

Study group (N = 12)

Control group (N = 8)

Mean

SD

Mean

Age Semantic fluency

29.75 62.83

5.82 12.97

32.37 66.38

6.80 7.15

Phonemic fluency

44.08

8.39

42.50

12.22

t (18)

p

−.92 −.70

.37 .49

SD

.34

.73

13



Exp Brain Res

Fig. 1  Experiment design

Table 2  Summary of mean percentage of error scores and response times (SD) in the Hayling test at T1

Active group

Sham group

p

Mean

SD

Initiation errors Initiation RTs Suppression error

3.33 606.87 20.17

3.62 152.04 11.97

4.17 582.82 21.87

1.05 111.95 10.52

−.60 .38 .33

.56 .71 .75

Suppression RTs

923.66

411.01

1300.35

407.32

−2.01

.06

(responses that were provided after the 3000 ms time limit), errors (words that are not compatible with the context of the sentence in the initiation condition; words that are compatible with the context of the sentence in the suppression condition), and the response times for each measure. Correct responses and error scores were converted to percentages. Voice response latencies were recorded using DirectRT software (v2012, Empirisoft). The response latencies were collected from the end of the final sentence word until the beginning of the voice response.

Mean

T SD

stimulations per week. At the end of the sixth stimulation, the participants repeated the Hayling test (T2). One month following the end of the stimulations, each participant completed the Hayling test for the third and last time (T3). Thus, all participants (in both the active and the sham groups) completed the same four lists of sentences at three time points: pre-stimulation (T1), immediately following the end of the stimulation sessions (T2, 2 weeks), and 1 month after T2 (T3). All four lists were used in each time point T1, T2, and T3. For each participant, the lists were presented in a randomized order, and the sentences in each list were randomized as well.

tDCS protocol tDCS was administered using a battery driven, constant current stimulator (neuroConn DC stimulator plus, Incl GmbH), and a pair of conductive rubber electrodes (7 cm × 5 cm, 35 cm2) covered in normal saline-soaked, synthetic sponge and restrained by a headband (Loo et al. 2011). Bilateral montage was conducted to increase the left DLPFC function and decrease the activation of the right DLPFC (Calautti et al. 2007; Kasahara et al. 2013). To stimulate the DLPFC, the anode electrode was placed over F3 (i.e., left DLPFC) and the cathode electrode over F4, i.e., right DLPFC), according to the 10–20 international system for EEG electrode placement. Each stimulation was applied for 20 min at 2 mA intensity (see Fig. 1). For the sham group, the stimulation ceased after 30 s. Procedure The study included seven meetings. At the first meeting, participants completed the phonemic and the semantic fluency tests, followed by the Hayling test. Immediately after performing the baseline (T1) Hayling test, the first stimulation session began. Each participant completed six sessions of 20 min of stimulation during 2 weeks, with three

13

Results Of the 20 participants tested, five reported experiencing mild itching during the active and the sham stimulation. No other adverse effects related to the application of tDCS were experience. First, we examined whether there were significant differences between the two study groups in the following measures: error scores and reaction times (RTs) of correct answers, in both the initiation and the suppression conditions before the stimulation (T1). A 2 × 3 repeated measures ANOVA was conducted on the error scores and the RTs with the active and sham groups as between-subject variable and time (T1, T2, and T3) as within-subject variable, separately for each condition (initiation and suppression). Table 2 shows that the groups did not differ in either error scores or RTs at baseline. Initiation condition Error scores To examine whether error scores in the initiation condition were reduced following the stimulation, a 2 × 3 repeated

Exp Brain Res Table 3  Mean percentage of error scores and RTs (ms) for correct answers (and SD), by group and time in the initiation condition Active Mean Error scores  T1 3.33  T2 2.08  T3 1.53 RTs for correct answers  T1 606.87  T2 569.70  T3

598.49

Table 4  Mean error scores and RTs in correct answers (and SD), by group and time in the suppression condition

Sham SD

Mean

Active SD

3.62 1.90 1.50

4.16 2.71 2.50

1.78 1.77 1.99

152.05 169.17

582.38 468.61

119.68 129.16

161.63

563.49

131.12

Mean Error scores  T1 20.17  T2 14.16  T3 12.64 RTs for correct responses  T1 923.66  T2 872.47  T3

870.26

Sham SD

Mean

SD

11.97 14.68 10.95

21.87 21.25 25.83

10.52 17.38 13.74

411.01 341.43

1300.35 1000.16

407.32 348.03

458.79

1112.98

378.71

measures ANOVA was conducted with the active and sham groups as between-subject variable and time (T1, T2, and T3) as within-subject variable. No main effects of group were found, F (1, 18) = 1.12, p = .30, η2 = .06, and time, F (2, 17) = 2.97, p = .08, η2 = .26 (Table 3). Moreover, the two-way interaction of group with time was insignificant, F (2, 17) = .07, p = .94, η2 = .01 (Table 3). Reaction times To examine improvement in response times (RTs) for correct answers in the initiation condition following the stimulation, a 2 × 3 repeated measures ANOVA was conducted with the active and sham groups as between-subject factor and the time (T1, T2, T3) as within-subject factor. No main effect of group was found, F (1, 18) = .75, p = .40, η2 = .04. A significant main effect of time, F (2, 17) = 6.26, p 

The effects of transcranial direct current stimulation over the dorsolateral prefrontal cortex on cognitive inhibition.

The present study examines the effects of bilateral transcranial direct current stimulation (tDCS) over the dorsolateral prefrontal cortex (DLPFC) (an...
564KB Sizes 0 Downloads 14 Views

Recommend Documents