- Email: [email protected]
Anodal tDCS of dorsolateral prefontal cortex during an Implicit Association Test
Neuroscience Letters 517 (2012) 82–86
Contents lists available at SciVerse ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Anodal tDCS of dorsolateral prefontal cortex during an Implicit Association Test Thomas E. Gladwin ∗ , Tess E. den Uyl, Reinout W. Wiers ADAPT Lab, Department of Psychology, University of Amsterdam, The Netherlands
a r t i c l e
i n f o
Article history: Received 11 February 2012 Received in revised form 24 March 2012 Accepted 7 April 2012 Keywords: tDCS IAT Executive function Bias Dorsolateral prefrontal cortex
a b s t r a c t Anodal stimulation of dorsolateral prefrontal cortex by transcranial Direct Current Stimulation (tDCS) has been shown to enhance performance on working memory tasks. However, it is not yet known precisely which aspects of working memory – a broad theoretical concept including short-term memory and various executive functions – are involved in such effects. In the current study, we aimed to determine whether tDCS would reduce bias effects on an Implicit Association Test, in which subjects must respond either congruently or incongruently to pre-existing evaluative associations. Such biases reflect a conflict between automatic associations and executive function, and tDCS was hypothesized to cause a shift in this balance in favor of executive function. The results clearly contradicted this hypothesis: tDCS did improve reaction times, but in the congruent rather than incongruent mapping condition. We conclude that DLPFC tDCS does not directly improve the ability to overcome bias; previous findings concerning working memory enhancement appear to reflect effects on a different component of executive function. © 2012 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Stimulation of the dorsolateral prefrontal cortex (DLPFC) with transcranial Direct Current Stimulation (tDCS) has been shown to improve performance on working memory tasks [2,3,10,17,22,28]. However, it remains unclear precisely how working memory – a broad concept related to short-term storage and manipulation of information [29], executive functions and reflective processes [11] – is affected by tDCS. In the current study, we focused on one important aspect of working memory: the ability to overcome biases resulting from the automatic activation of existing evaluative associations [8,26]. This function plays a role in the Implicit Association Test, IAT [13], a widely used method to detect associations between target concepts and evaluative attributes [14]. For example, a target concept could be spiders, and the evaluative attribute fear. The test requires subjects to classify target and attribute stimuli into one of two categories via key presses. In the essential blocks, subjects must respond to both target and attribute stimuli. Each response key is mapped to one target category and one attribute category. This mapping is either congruent (e.g. spider and fear) or incongruent (e.g. spider and tasty) to the subject’s evaluative associations. Subjects show better performance when this mapping is congruent [13,21].
∗ Corresponding author at: Weesperplein 4, Room 6.15, 1018 XB Amsterdam, The Netherlands. Tel.: +31 06 18091496; fax: +31 20 5256710. E-mail addresses: [email protected], [email protected] (T.E. Gladwin). 0304-3940/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.neulet.2012.04.025
The bias defined by the difference in performance on incongruent versus congruent blocks has been argued, supported by mathematical modeling, to reflect a mixture of automatic and controlled processes [8]. Automatic associations will tend to cause incorrect responses in incongruent blocks, which may be overcome given sufficient cognitive control. Indeed, performing incongruent versus congruent blocks evokes activation in prefrontal brain regions including left DLPFC [6,18,19], a region associated with working memory processes [1,5,23,24,27]. In line with these results, and further establishing a causal role for DLPFC, TMS applied to left DLPFC has been found to increase bias on the IAT [4], due to increased errors in incongruent trials; no effects on reaction time (RT) were found. This leads to the expectation that that bias would be reduced by anodal stimulation of left DLPFC: if this region is involved with the ability to overcome the effects of automatic processes, the enhancement of working memory caused by its stimulation should lead to fewer errors or faster responses in incongruent blocks. We therefore hypothesized that anodal prefrontal stimulation during a classic insects/flowers IAT, would reduce bias on the IAT. 2. Materials and methods Twenty Dutch speaking students or ex-students were tested (13 female, mean age = 21.1, SD = 2.5). All participants gave informed consent and the study was approved by the local ethics committee. Subjects received money (10 euros) for participation. TDCS was administered via two electrodes with saline soaked 5 × 7 sponges that were placed on the scalp. The anodal electrode was placed over DLPFC, at the F3 position in the 10–20 EEG system. The cathodal
T.E. Gladwin et al. / Neuroscience Letters 517 (2012) 82–86
electrode was placed over the right orbit. During active tDCS stimulation participants were stimulated with 1 mA for 10 min preceding task performance; during sham stimulation the stimulation was automatically switched off after 1 min, without the subject being informed of this. Following the 10 min of active or sham stimulation, participants rated eight side effects (itching, burning, tingling, pain, headache, vertigo, nausea, fatigue) on a 0 (none) to 4 (strong) scale. The study used a within-subject design. Each participant received sham stimulation and active stimulation on two testing days, in counterbalanced order. There was at least one day and no more than eight days between the two sessions (M = 4.7, SD = 2.5). Following stimulation, subjects performed the task, which they had briefly practiced beforehand. The task was a variant of the classic insect/flower IAT [13], shown in Fig. 1. In this task, the target categories are insects (e.g., cockroach, moth) and flowers (e.g., sunflower, rose), and the attribute categories are positive (e.g., friends, happy) versus negative (e.g., filth, disease). The task is known to evoke a strong bias: subjects find it easier to classify insects together with the negative category and flower with the positive category. Subjects responded using left (the F-key) and right (the J-key) response keys. The category labels currently mapped to the left and right response keys were presented on the corresponding sides of the screen. On each trial, a word was presented to be categorized via button press. The task consisted of three repetitions of a sequence of seven blocks. The first blocks of each sequence required only target words to be categorized; the second block also required only target words to be categorized, but reversed the mapping of target categories to the left and right keys. The third block required only attribute stimuli to be categorized. Subsequently four mixed blocks were presented, in which both target and attribute categories were mapped to the response keys. The four mixed blocks alternated between congruent and incongruent blocks, with counterbalanced starting block type. In congruent blocks, negative and insect words were mapped to one key, and positive and flower words to the other key; in incongruent blocks, negative and flower
83
words were mapped to one key, and positive and insect words to the other key. The mapping of attribute words to left and right keys was counterbalanced over subjects. Reaction times and accuracy on the mixed blocks were analyzed using a repeated measures ANOVA. Independent variables were tDCS stimulation (sham versus active), block type (congruent versus incongruent block) and stimulus type (target versus attribute). In follow-up analyses, effects of tDCS were tested for each block type.
3. Results Overall the tDCS procedure was well received and few side effects were reported. Itching and burning sensations were most often reported, but there was no difference between sham and active stimulation (sham: 0.52, active: 0.51). When asked to guess which session was the active stimulation, 12 of the 20 subjects guessed correctly, i.e., at close to chance level. Almost all participants reported believing the stimulation in the first session to be the active stimulation (17 out of 20). On RT (Tables 1 and 2, Fig. 2), effects were found, first, of block type: responses were slower on incongruent blocks than on congruent blocks (F(1, 19) = 50.40, p < 0.0001, 2p = 0.73). Responses to target stimuli were faster than to attribute stimuli (F(1, 19) = 5.22, p = 0.034, 2p = 0.22). An interaction between active versus sham stimulation, stimulus type and block type was found (F(1, 19) = 10.60, p = 0.0042, 2p = 0.36). Pairwise comparisons showed that this was due to an improvement of RT by tDCS specifically for targets in congruent blocks (t(19) = −2.91, p = 0.0090). On accuracy (Tables 1 and 2, Fig. 2), more errors were found in incongruent blocks than in congruent blocks (F(1, 19) = 91.39, p < 0.0001, 2p = 0.83). Responses to targets were less accurate than to attribute words (F(1, 19) = 35.60, p < 0.0001, 2p = 0.65), and this effect was marginally significantly larger in congruent than in incongruent blocks (F(1, 19) = 4.38, p = 0.050, 2p = 0.19).
Fig. 1. Illustration of the Implicit Association Task. (A) An example of three trials from a congruent block. Subjects must press the left key for both Insect words and Bad words, and the right key for both Flower words and Good words. In the first trial of the congruent block, the subject sees “Lily” and should therefore press the right (Flower/Good) key. Response keys are represented by circles; correct responses are indicated by the filled circle. (B) An example of three trials from an incongruent block. The left key is now mapped to Insect and Good words, and the right key to Flower and Bad words. In the first trial, the correct response to the stimulus “Spider” would therefore be the Insect/Good key.
84
T.E. Gladwin et al. / Neuroscience Letters 517 (2012) 82–86
Table 1 Descriptive statistics for performance on the IAT. Sham tDCS
Active tDCS
Congruent
RT Acc
Incongruent
Congruent
Incongruent
T
A
T
A
T
A
T
A
774.2 (23) 0.94 (0.01)
874.2 (26) 0.88 (0.011)
782.1 (24) 0.97 (0.0057)
931.9 (36) 0.94 (0.0097)
715.4 (19) 0.94 (0.0097)
870.4 (30) 0.88 (0.012)
744.6 (25) 0.97 (0.0053)
890.5 (37) 0.94 (0.012)
Note. Mean reaction time [ms] and accuracy [proportion correct]; standard errors are printed in parentheses.
4. Discussion The results contradicted the hypothesis. TDCS did significantly improve rather than disrupt performance, but this effect was found in congruent blocks rather than in incongruent blocks. Thus, DLFC tDCS does not appear to directly improve the ability to overcome bias. This was unexpected given the usual interpretation of prior research: broadly, that DLPFC is involved with overcoming bias [4,6,18], and that the tDCS method used in the current study facilitates processing in DLPFC and thus enhances working memory [2,3,10,17,22,28]. However, in a recent study we already attempted to nuance the interpretation of the effect of tDCS on working memory [12]. Briefly, subjects performed a Sternberg task in which distractor stimuli were presented during the retention period. TDCS was found to improve RT, but only when subjects had to distinguish between distractors and items from the memory set. It appeared that a specific component of working memory
was improved by anodal DLPFC stimulation, such that relationships between stimuli and their behavioral relevance were more readily accessible. Responding in the incongruent condition of an Implicit Association Test, which requires the overcoming of a response bias, is strongly associated with worse performance in terms of reaction time and accuracy, and this basic finding was replicated in the current study. The stimulation of DLPFC did significantly affect performance, but did not improve subjects’ ability to overcome bias. Thus the question is which component of working memory could be improved by DLPFC stimulation, if not overcoming bias? The current results cannot distinguish between various, possibly non-exclusive possibilities, e.g., working memory monitoring [5], the facilitation of relevant information (as opposed to the inhibition of irrelevant information) [9], the detection of mismatch relative to the content of working memory [30] or the maintenance of task goals [7]. However, considering the current task, one
Fig. 2. Reaction times and accuracy. Performance on congruent and incongruent blocks, for target and attribute stimulus types, after sham and active tDCS. Bars show mean values per condition; error bars show standard errors. (A) Reaction times in milliseconds. (B) Accuracy in proportion correct.
T.E. Gladwin et al. / Neuroscience Letters 517 (2012) 82–86 Table 2 Overview of results. Measure
Effect
F(1, 19)
P
2p
Reaction time
Incongruence Stimulus type Stimulation × stimulus type × incongruence Incongruence Stimulus type Stimulus type × incongruence
50.40 5.22 10.60
<0.0001 0.034 0.0042
0.73 0.22 0.36
91.39 35.60 4.38
<0.0001 <0.0001 0.050
0.83 0.65 0.19
Accuracy
Note. Overview of results found on reaction time and accuracy. Incongruence: incongruent versus congruent block. Stimulus type: target versus attribute word. Stimulation: active versus sham tDCS.
tentative explanation of the results in line with previous research is that that stimulation activated all currently relevant relationships between stimuli and (behavioral or attentional) responses – essentially, both those based on long-term memory, and those based on current task instructions. Such associations could be described in terms of working memory monitoring or the amplification of information relevant to the current situation. In congruent blocks, the enhanced activation of associations would lead to faster reaction times as expected. However, in incongruent blocks, simply enhancing the activation of associations would equally affect both congruent and incongruent associations. Importantly, their competition would not necessarily be resolved more quickly or accurately due to such indiscriminate enhancement at the level of DLPFC. Note that this interpretation does not contradict previous findings on overcoming bias involving activation of DLPFC and disruption of DLPFC by TMS. For weaker associations to compete with preexisting associations could well require increased DLPFC activation, and for this process to be resolved in favor of the weaker association could be more sensitive to disruption. This does not imply that DLPFC stimulation would necessarily selectively favor the activation of incongruent associations. Further, DLPFC may be related to a sub-component of executive function that is necessary for, but not identifiable with, overcoming bias or inhibitory control [15]. The results may thus help to further specify the precise effects of DLPFC stimulation on cognition, and hence also help understand the precise function of DLPFC itself. Future research is needed to test the current interpretation of the effect of DLPFC stimulation in terms of component processes of working memory. Another unexpected effect in need of further study was the dependence of the effect of tDCS on stimulus type, suggesting that associations differ between targets and attributes. This may reflect the one-tomany relationship between attributes and targets; e.g., an object about which a subject is anxious may strongly evoke fear, while the concept of fear may evoke a more diffuse set of feared targets. Further, the current study has limitations that further work should address. First, only anodal and sham stimulation at position F3 was used, but no cathodal stimulation or stimulation of alternative sites (see, e.g., [16]). It therefore remains to be determined how specific the current results are to the stimulation parameters; e.g., whether anodal stimulation of left motor cortex or parietal cortex would cause equivalent or different effects, or whether cathodal, i.e., inhibitory stimulation would result in different or null effects. In conclusion, it has been established in multiple studies that prefrontal stimulation using tDCS can enhance aspects of working memory; the question is which specific underlying processes are involved. The current results did not support the hypothesis that this process involves overcoming bias. Taken together, the current study and recent results [12] appear to provide tentative support to an interpretation of DLPFC function in terms of encoding what needs to be done with information in working memory,
85
rather than actually executing an associated cognitive function such as overcoming bias. Finally, we briefly note that evaluative associations, as measured by the IAT, play a role in various forms of psychopathology [25]. The current results may therefore be of relevance to studies aimed at clinical applications of tDCS, especially given known effects of tDCS on neuroplasticity [20]. Further study of effects of tDCS on the processes underlying such associations may therefore have significant theoretical and practical implications. Acknowledgments The authors are supported by VICI award 453.08.01 from the Netherlands National Science Foundation (N.W.O.) and FMG-UvA Research Priority Grant on Affect-Regulation. References [1] C. Amiez, M. Petrides, Selective involvement of the mid-dorsolateral prefrontal cortex in the coding of the serial order of visual stimuli in working memory, Proceedings of the National Academy of Sciences of the United States of America 104 (2007) 13786–13791. [2] P.S. Boggio, R. Ferrucci, S.P. Rigonatti, P. Covre, M. Nitsche, A. Pascual-Leone, F. Fregni, Effects of transcranial direct current stimulation on working memory in patients with Parkinson’s disease, Journal of the Neurological Sciences 249 (2006) 31–38. [3] P.S. Boggio, L.P. Khoury, D.C.S. Martins, O.E.M.S. Martins, E.C. de Macedo, F. Fregni, Temporal cortex direct current stimulation enhances performance on a visual recognition memory task in Alzheimer disease, Journal of Neurology, Neurosurgery and Psychiatry 80 (2009) 444–447. [4] Z. Cattaneo, G. Mattavelli, E. Platania, C. Papagno, The role of the prefrontal cortex in controlling gender-stereotypical associations: a TMS investigation, NeuroImage 56 (2011) 1839–1846. [5] A.S. Champod, M. Petrides, Dissociable roles of the posterior parietal and the prefrontal cortex in manipulation and monitoring processes, Proceedings of the National Academy of Sciences of the United States of America 104 (2007) 14837–14842. [6] M.W. Chee, N. Sriram, C.S. Soon, K.M. Lee, Dorsolateral prefrontal cortex and the implicit association of concepts and attributes, Neuroreport 11 (2000) 135–140. [7] M.W. Cole, W. Schneider, The cognitive control network: integrated cortical regions with dissociable functions, NeuroImage 37 (2007) 343–360. [8] F.R. Conrey, J.W. Sherman, B. Gawronski, K. Hugenberg, C.J. Groom, Separating multiple processes in implicit social cognition: the quad model of implicit task performance, Journal of Personality and Social Psychology 89 (2005) 469–487. [9] T. Egner, J. Hirsch, Cognitive control mechanisms resolve conflict through cortical amplification of task-relevant information, Nature Neuroscience 8 (2005) 1784–1790. [10] F. Fregni, P.S. Boggio, M. Nitsche, F. Bermpohl, A. Antal, E. Feredoes, M.A. Marcolin, S.P. Rigonatti, M.T. Silva, W. Paulus, A. Pascual-Leone, Anodal transcranial direct current stimulation of prefrontal cortex enhances working memory, Experimental Brain Research 166 (2005) 23–30. [11] T.E. Gladwin, B. Figner, E.A. Crone, R.W. Wiers, Addiction, adolescence, and the integration of control and motivation, Developmental Cognitive Neuroscience (2011) 364–376. [12] T.E. Gladwin, T. den Uyl, F.F. Fregni, R.W. Wiers, Enhancement of selective attention by tDCS: interaction with interference in a Sternberg task, Neuroscience Letters (2012) http://dx.doi.org/10.1016/j.neulet.2012.01.056. [13] A.G. Greenwald, D.E. McGhee, J.L. Schwartz, Measuring individual differences in implicit cognition: the implicit association test, Journal of Personality and Social Psychology 74 (1998) 1464–1480. [14] A.G. Greenwald, T.A. Poehlman, E.L. Uhlmann, M.R. Banaji, Understanding and using the Implicit Association Test: III. Meta-analysis of predictive validity, Journal of Personality and Social Psychology 97 (2009) 17–41. [15] R. Hester, K. Murphy, H. Garavan, Beyond common resources: the cortical basis for resolving task interference, NeuroImage 23 (2004) 202–212. [16] A.H. Javadi, V. Walsh, Transcranial direct current stimulation (tDCS) of the left dorsolateral prefrontal cortex modulates declarative memory, Brain Stimulation 12 (2) (2011), http://dx.doi.org/10.1186/1471-2202-12-2. [17] J.M. Jo, Y.-H. Kim, M.-H. Ko, S.H. Ohn, B. Joen, K.H. Lee, Enhancing the working memory of stroke patients using tDCS, American Journal of Physical Medicine and Rehabilitation 88 (2009) 404–409. [18] K.M. Knutson, L. Mah, C.F. Manly, J. Grafman, Neural correlates of automatic beliefs about gender and race, Human Brain Mapping 28 (2007) 915–930. [19] Q. Luo, M. Nakic, T. Wheatley, R. Richell, A. Martin, R.J.R. Blair, The neural basis of implicit moral attitude—an IAT study using event-related fMRI, NeuroImage 30 (2006) 1449–1457. [20] M.A. Nitsche, A. Roth, M.F. Kuo, A.K. Fischer, D. Liebetanz, N. Lang, F. Tergau, W. Paulus, Timing-dependent modulation of associative plasticity by general network excitability in the human motor cortex, Neuroscience 27 (2007) 3807–3812.
86
T.E. Gladwin et al. / Neuroscience Letters 517 (2012) 82–86
[21] B.A. Nosek, A.G. Greenwald, M.R. Banaji, Understanding and using the Implicit Association Test: II. Method variables and construct validity, Personality and Social Psychology Bulletin 31 (2005) 166–180. [22] S.H. Ohn, C.I. Park, W.K. Yoo, M.H. Ko, K.P. Choi, G.M. Kim, Y.T. Lee, Y.H. Lee, Timedependent effect of transcranial direct current stimulation on the enhancement of working memory, Neuroreport 19 (2008) 43–47. [23] A.M. Owen, N.J. Herrod, D.K. Menon, J.C. Clark, S.P. Downey, T.A. Carpenter, P.S. Minhas, F.E. Turkheimer, E.J. Williams, T.W. Robbins, B.J. Sahakian, M. Petrides, J.D. Pickard, Redefining the functional organization of working memory processes within human lateral prefrontal cortex, European Journal of Neuroscience 11 (1999) 567–574. [24] M. Petrides, The role of the mid-dorsolateral prefrontal cortex in working memory, Experimental Brain Research 133 (2000) 44–54. [25] A. Roefs, J. Huijding, F.T.Y. Smulders, C.M. MacLeod, P.J. de Jong, R.W. Wiers, A.T. Jansen, Implicit measures of association in psychopathology research, Psychological Bulletin 137 (2011) 149–193.
[26] J.W. Sherman, B. Gawronski, K. Gonsalkorale, K. Hugenberg, T.J. Allen, C.J. Groom, The self-regulation of automatic associations and behavioral impulses, Psychological Review 115 (2008) 314–335. [27] E.E. Smith, J. Jonides, Neuroimaging analyses of human working memory, Proceedings of the National Academy of Sciences of the United States of America 95 (1998) 12061–12068. [28] R. Sparing, M. Dafotakis, I.G. Meister, N. Thirugnanasambandam, G.R. Fink, Enhancing language performance with non-invasive brain stimulation—a transcranial direct current stimulation study in healthy humans, Neuropsychologia 46 (2008) 261–268. [29] N. Unsworth, R.W. Engle, On the division of short-term and working memory: an examination of simple and complex span and their relation to higher order abilities, Psychological Bulletin 133 (2007) 1038–1066. [30] X. Zhang, L. Ma, S. Li, Y. Wang, X. Weng, L. Wang, A mismatch process in brief delayed matching-to-sample task: an fMRI study, Experimental Brain Research 186 (2008) 335–341.
Contents lists available at SciVerse ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Anodal tDCS of dorsolateral prefontal cortex during an Implicit Association Test Thomas E. Gladwin ∗ , Tess E. den Uyl, Reinout W. Wiers ADAPT Lab, Department of Psychology, University of Amsterdam, The Netherlands
a r t i c l e
i n f o
Article history: Received 11 February 2012 Received in revised form 24 March 2012 Accepted 7 April 2012 Keywords: tDCS IAT Executive function Bias Dorsolateral prefrontal cortex
a b s t r a c t Anodal stimulation of dorsolateral prefrontal cortex by transcranial Direct Current Stimulation (tDCS) has been shown to enhance performance on working memory tasks. However, it is not yet known precisely which aspects of working memory – a broad theoretical concept including short-term memory and various executive functions – are involved in such effects. In the current study, we aimed to determine whether tDCS would reduce bias effects on an Implicit Association Test, in which subjects must respond either congruently or incongruently to pre-existing evaluative associations. Such biases reflect a conflict between automatic associations and executive function, and tDCS was hypothesized to cause a shift in this balance in favor of executive function. The results clearly contradicted this hypothesis: tDCS did improve reaction times, but in the congruent rather than incongruent mapping condition. We conclude that DLPFC tDCS does not directly improve the ability to overcome bias; previous findings concerning working memory enhancement appear to reflect effects on a different component of executive function. © 2012 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Stimulation of the dorsolateral prefrontal cortex (DLPFC) with transcranial Direct Current Stimulation (tDCS) has been shown to improve performance on working memory tasks [2,3,10,17,22,28]. However, it remains unclear precisely how working memory – a broad concept related to short-term storage and manipulation of information [29], executive functions and reflective processes [11] – is affected by tDCS. In the current study, we focused on one important aspect of working memory: the ability to overcome biases resulting from the automatic activation of existing evaluative associations [8,26]. This function plays a role in the Implicit Association Test, IAT [13], a widely used method to detect associations between target concepts and evaluative attributes [14]. For example, a target concept could be spiders, and the evaluative attribute fear. The test requires subjects to classify target and attribute stimuli into one of two categories via key presses. In the essential blocks, subjects must respond to both target and attribute stimuli. Each response key is mapped to one target category and one attribute category. This mapping is either congruent (e.g. spider and fear) or incongruent (e.g. spider and tasty) to the subject’s evaluative associations. Subjects show better performance when this mapping is congruent [13,21].
∗ Corresponding author at: Weesperplein 4, Room 6.15, 1018 XB Amsterdam, The Netherlands. Tel.: +31 06 18091496; fax: +31 20 5256710. E-mail addresses: [email protected], [email protected] (T.E. Gladwin). 0304-3940/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.neulet.2012.04.025
The bias defined by the difference in performance on incongruent versus congruent blocks has been argued, supported by mathematical modeling, to reflect a mixture of automatic and controlled processes [8]. Automatic associations will tend to cause incorrect responses in incongruent blocks, which may be overcome given sufficient cognitive control. Indeed, performing incongruent versus congruent blocks evokes activation in prefrontal brain regions including left DLPFC [6,18,19], a region associated with working memory processes [1,5,23,24,27]. In line with these results, and further establishing a causal role for DLPFC, TMS applied to left DLPFC has been found to increase bias on the IAT [4], due to increased errors in incongruent trials; no effects on reaction time (RT) were found. This leads to the expectation that that bias would be reduced by anodal stimulation of left DLPFC: if this region is involved with the ability to overcome the effects of automatic processes, the enhancement of working memory caused by its stimulation should lead to fewer errors or faster responses in incongruent blocks. We therefore hypothesized that anodal prefrontal stimulation during a classic insects/flowers IAT, would reduce bias on the IAT. 2. Materials and methods Twenty Dutch speaking students or ex-students were tested (13 female, mean age = 21.1, SD = 2.5). All participants gave informed consent and the study was approved by the local ethics committee. Subjects received money (10 euros) for participation. TDCS was administered via two electrodes with saline soaked 5 × 7 sponges that were placed on the scalp. The anodal electrode was placed over DLPFC, at the F3 position in the 10–20 EEG system. The cathodal
T.E. Gladwin et al. / Neuroscience Letters 517 (2012) 82–86
electrode was placed over the right orbit. During active tDCS stimulation participants were stimulated with 1 mA for 10 min preceding task performance; during sham stimulation the stimulation was automatically switched off after 1 min, without the subject being informed of this. Following the 10 min of active or sham stimulation, participants rated eight side effects (itching, burning, tingling, pain, headache, vertigo, nausea, fatigue) on a 0 (none) to 4 (strong) scale. The study used a within-subject design. Each participant received sham stimulation and active stimulation on two testing days, in counterbalanced order. There was at least one day and no more than eight days between the two sessions (M = 4.7, SD = 2.5). Following stimulation, subjects performed the task, which they had briefly practiced beforehand. The task was a variant of the classic insect/flower IAT [13], shown in Fig. 1. In this task, the target categories are insects (e.g., cockroach, moth) and flowers (e.g., sunflower, rose), and the attribute categories are positive (e.g., friends, happy) versus negative (e.g., filth, disease). The task is known to evoke a strong bias: subjects find it easier to classify insects together with the negative category and flower with the positive category. Subjects responded using left (the F-key) and right (the J-key) response keys. The category labels currently mapped to the left and right response keys were presented on the corresponding sides of the screen. On each trial, a word was presented to be categorized via button press. The task consisted of three repetitions of a sequence of seven blocks. The first blocks of each sequence required only target words to be categorized; the second block also required only target words to be categorized, but reversed the mapping of target categories to the left and right keys. The third block required only attribute stimuli to be categorized. Subsequently four mixed blocks were presented, in which both target and attribute categories were mapped to the response keys. The four mixed blocks alternated between congruent and incongruent blocks, with counterbalanced starting block type. In congruent blocks, negative and insect words were mapped to one key, and positive and flower words to the other key; in incongruent blocks, negative and flower
83
words were mapped to one key, and positive and insect words to the other key. The mapping of attribute words to left and right keys was counterbalanced over subjects. Reaction times and accuracy on the mixed blocks were analyzed using a repeated measures ANOVA. Independent variables were tDCS stimulation (sham versus active), block type (congruent versus incongruent block) and stimulus type (target versus attribute). In follow-up analyses, effects of tDCS were tested for each block type.
3. Results Overall the tDCS procedure was well received and few side effects were reported. Itching and burning sensations were most often reported, but there was no difference between sham and active stimulation (sham: 0.52, active: 0.51). When asked to guess which session was the active stimulation, 12 of the 20 subjects guessed correctly, i.e., at close to chance level. Almost all participants reported believing the stimulation in the first session to be the active stimulation (17 out of 20). On RT (Tables 1 and 2, Fig. 2), effects were found, first, of block type: responses were slower on incongruent blocks than on congruent blocks (F(1, 19) = 50.40, p < 0.0001, 2p = 0.73). Responses to target stimuli were faster than to attribute stimuli (F(1, 19) = 5.22, p = 0.034, 2p = 0.22). An interaction between active versus sham stimulation, stimulus type and block type was found (F(1, 19) = 10.60, p = 0.0042, 2p = 0.36). Pairwise comparisons showed that this was due to an improvement of RT by tDCS specifically for targets in congruent blocks (t(19) = −2.91, p = 0.0090). On accuracy (Tables 1 and 2, Fig. 2), more errors were found in incongruent blocks than in congruent blocks (F(1, 19) = 91.39, p < 0.0001, 2p = 0.83). Responses to targets were less accurate than to attribute words (F(1, 19) = 35.60, p < 0.0001, 2p = 0.65), and this effect was marginally significantly larger in congruent than in incongruent blocks (F(1, 19) = 4.38, p = 0.050, 2p = 0.19).
Fig. 1. Illustration of the Implicit Association Task. (A) An example of three trials from a congruent block. Subjects must press the left key for both Insect words and Bad words, and the right key for both Flower words and Good words. In the first trial of the congruent block, the subject sees “Lily” and should therefore press the right (Flower/Good) key. Response keys are represented by circles; correct responses are indicated by the filled circle. (B) An example of three trials from an incongruent block. The left key is now mapped to Insect and Good words, and the right key to Flower and Bad words. In the first trial, the correct response to the stimulus “Spider” would therefore be the Insect/Good key.
84
T.E. Gladwin et al. / Neuroscience Letters 517 (2012) 82–86
Table 1 Descriptive statistics for performance on the IAT. Sham tDCS
Active tDCS
Congruent
RT Acc
Incongruent
Congruent
Incongruent
T
A
T
A
T
A
T
A
774.2 (23) 0.94 (0.01)
874.2 (26) 0.88 (0.011)
782.1 (24) 0.97 (0.0057)
931.9 (36) 0.94 (0.0097)
715.4 (19) 0.94 (0.0097)
870.4 (30) 0.88 (0.012)
744.6 (25) 0.97 (0.0053)
890.5 (37) 0.94 (0.012)
Note. Mean reaction time [ms] and accuracy [proportion correct]; standard errors are printed in parentheses.
4. Discussion The results contradicted the hypothesis. TDCS did significantly improve rather than disrupt performance, but this effect was found in congruent blocks rather than in incongruent blocks. Thus, DLFC tDCS does not appear to directly improve the ability to overcome bias. This was unexpected given the usual interpretation of prior research: broadly, that DLPFC is involved with overcoming bias [4,6,18], and that the tDCS method used in the current study facilitates processing in DLPFC and thus enhances working memory [2,3,10,17,22,28]. However, in a recent study we already attempted to nuance the interpretation of the effect of tDCS on working memory [12]. Briefly, subjects performed a Sternberg task in which distractor stimuli were presented during the retention period. TDCS was found to improve RT, but only when subjects had to distinguish between distractors and items from the memory set. It appeared that a specific component of working memory
was improved by anodal DLPFC stimulation, such that relationships between stimuli and their behavioral relevance were more readily accessible. Responding in the incongruent condition of an Implicit Association Test, which requires the overcoming of a response bias, is strongly associated with worse performance in terms of reaction time and accuracy, and this basic finding was replicated in the current study. The stimulation of DLPFC did significantly affect performance, but did not improve subjects’ ability to overcome bias. Thus the question is which component of working memory could be improved by DLPFC stimulation, if not overcoming bias? The current results cannot distinguish between various, possibly non-exclusive possibilities, e.g., working memory monitoring [5], the facilitation of relevant information (as opposed to the inhibition of irrelevant information) [9], the detection of mismatch relative to the content of working memory [30] or the maintenance of task goals [7]. However, considering the current task, one
Fig. 2. Reaction times and accuracy. Performance on congruent and incongruent blocks, for target and attribute stimulus types, after sham and active tDCS. Bars show mean values per condition; error bars show standard errors. (A) Reaction times in milliseconds. (B) Accuracy in proportion correct.
T.E. Gladwin et al. / Neuroscience Letters 517 (2012) 82–86 Table 2 Overview of results. Measure
Effect
F(1, 19)
P
2p
Reaction time
Incongruence Stimulus type Stimulation × stimulus type × incongruence Incongruence Stimulus type Stimulus type × incongruence
50.40 5.22 10.60
<0.0001 0.034 0.0042
0.73 0.22 0.36
91.39 35.60 4.38
<0.0001 <0.0001 0.050
0.83 0.65 0.19
Accuracy
Note. Overview of results found on reaction time and accuracy. Incongruence: incongruent versus congruent block. Stimulus type: target versus attribute word. Stimulation: active versus sham tDCS.
tentative explanation of the results in line with previous research is that that stimulation activated all currently relevant relationships between stimuli and (behavioral or attentional) responses – essentially, both those based on long-term memory, and those based on current task instructions. Such associations could be described in terms of working memory monitoring or the amplification of information relevant to the current situation. In congruent blocks, the enhanced activation of associations would lead to faster reaction times as expected. However, in incongruent blocks, simply enhancing the activation of associations would equally affect both congruent and incongruent associations. Importantly, their competition would not necessarily be resolved more quickly or accurately due to such indiscriminate enhancement at the level of DLPFC. Note that this interpretation does not contradict previous findings on overcoming bias involving activation of DLPFC and disruption of DLPFC by TMS. For weaker associations to compete with preexisting associations could well require increased DLPFC activation, and for this process to be resolved in favor of the weaker association could be more sensitive to disruption. This does not imply that DLPFC stimulation would necessarily selectively favor the activation of incongruent associations. Further, DLPFC may be related to a sub-component of executive function that is necessary for, but not identifiable with, overcoming bias or inhibitory control [15]. The results may thus help to further specify the precise effects of DLPFC stimulation on cognition, and hence also help understand the precise function of DLPFC itself. Future research is needed to test the current interpretation of the effect of DLPFC stimulation in terms of component processes of working memory. Another unexpected effect in need of further study was the dependence of the effect of tDCS on stimulus type, suggesting that associations differ between targets and attributes. This may reflect the one-tomany relationship between attributes and targets; e.g., an object about which a subject is anxious may strongly evoke fear, while the concept of fear may evoke a more diffuse set of feared targets. Further, the current study has limitations that further work should address. First, only anodal and sham stimulation at position F3 was used, but no cathodal stimulation or stimulation of alternative sites (see, e.g., [16]). It therefore remains to be determined how specific the current results are to the stimulation parameters; e.g., whether anodal stimulation of left motor cortex or parietal cortex would cause equivalent or different effects, or whether cathodal, i.e., inhibitory stimulation would result in different or null effects. In conclusion, it has been established in multiple studies that prefrontal stimulation using tDCS can enhance aspects of working memory; the question is which specific underlying processes are involved. The current results did not support the hypothesis that this process involves overcoming bias. Taken together, the current study and recent results [12] appear to provide tentative support to an interpretation of DLPFC function in terms of encoding what needs to be done with information in working memory,
85
rather than actually executing an associated cognitive function such as overcoming bias. Finally, we briefly note that evaluative associations, as measured by the IAT, play a role in various forms of psychopathology [25]. The current results may therefore be of relevance to studies aimed at clinical applications of tDCS, especially given known effects of tDCS on neuroplasticity [20]. Further study of effects of tDCS on the processes underlying such associations may therefore have significant theoretical and practical implications. Acknowledgments The authors are supported by VICI award 453.08.01 from the Netherlands National Science Foundation (N.W.O.) and FMG-UvA Research Priority Grant on Affect-Regulation. References [1] C. Amiez, M. Petrides, Selective involvement of the mid-dorsolateral prefrontal cortex in the coding of the serial order of visual stimuli in working memory, Proceedings of the National Academy of Sciences of the United States of America 104 (2007) 13786–13791. [2] P.S. Boggio, R. Ferrucci, S.P. Rigonatti, P. Covre, M. Nitsche, A. Pascual-Leone, F. Fregni, Effects of transcranial direct current stimulation on working memory in patients with Parkinson’s disease, Journal of the Neurological Sciences 249 (2006) 31–38. [3] P.S. Boggio, L.P. Khoury, D.C.S. Martins, O.E.M.S. Martins, E.C. de Macedo, F. Fregni, Temporal cortex direct current stimulation enhances performance on a visual recognition memory task in Alzheimer disease, Journal of Neurology, Neurosurgery and Psychiatry 80 (2009) 444–447. [4] Z. Cattaneo, G. Mattavelli, E. Platania, C. Papagno, The role of the prefrontal cortex in controlling gender-stereotypical associations: a TMS investigation, NeuroImage 56 (2011) 1839–1846. [5] A.S. Champod, M. Petrides, Dissociable roles of the posterior parietal and the prefrontal cortex in manipulation and monitoring processes, Proceedings of the National Academy of Sciences of the United States of America 104 (2007) 14837–14842. [6] M.W. Chee, N. Sriram, C.S. Soon, K.M. Lee, Dorsolateral prefrontal cortex and the implicit association of concepts and attributes, Neuroreport 11 (2000) 135–140. [7] M.W. Cole, W. Schneider, The cognitive control network: integrated cortical regions with dissociable functions, NeuroImage 37 (2007) 343–360. [8] F.R. Conrey, J.W. Sherman, B. Gawronski, K. Hugenberg, C.J. Groom, Separating multiple processes in implicit social cognition: the quad model of implicit task performance, Journal of Personality and Social Psychology 89 (2005) 469–487. [9] T. Egner, J. Hirsch, Cognitive control mechanisms resolve conflict through cortical amplification of task-relevant information, Nature Neuroscience 8 (2005) 1784–1790. [10] F. Fregni, P.S. Boggio, M. Nitsche, F. Bermpohl, A. Antal, E. Feredoes, M.A. Marcolin, S.P. Rigonatti, M.T. Silva, W. Paulus, A. Pascual-Leone, Anodal transcranial direct current stimulation of prefrontal cortex enhances working memory, Experimental Brain Research 166 (2005) 23–30. [11] T.E. Gladwin, B. Figner, E.A. Crone, R.W. Wiers, Addiction, adolescence, and the integration of control and motivation, Developmental Cognitive Neuroscience (2011) 364–376. [12] T.E. Gladwin, T. den Uyl, F.F. Fregni, R.W. Wiers, Enhancement of selective attention by tDCS: interaction with interference in a Sternberg task, Neuroscience Letters (2012) http://dx.doi.org/10.1016/j.neulet.2012.01.056. [13] A.G. Greenwald, D.E. McGhee, J.L. Schwartz, Measuring individual differences in implicit cognition: the implicit association test, Journal of Personality and Social Psychology 74 (1998) 1464–1480. [14] A.G. Greenwald, T.A. Poehlman, E.L. Uhlmann, M.R. Banaji, Understanding and using the Implicit Association Test: III. Meta-analysis of predictive validity, Journal of Personality and Social Psychology 97 (2009) 17–41. [15] R. Hester, K. Murphy, H. Garavan, Beyond common resources: the cortical basis for resolving task interference, NeuroImage 23 (2004) 202–212. [16] A.H. Javadi, V. Walsh, Transcranial direct current stimulation (tDCS) of the left dorsolateral prefrontal cortex modulates declarative memory, Brain Stimulation 12 (2) (2011), http://dx.doi.org/10.1186/1471-2202-12-2. [17] J.M. Jo, Y.-H. Kim, M.-H. Ko, S.H. Ohn, B. Joen, K.H. Lee, Enhancing the working memory of stroke patients using tDCS, American Journal of Physical Medicine and Rehabilitation 88 (2009) 404–409. [18] K.M. Knutson, L. Mah, C.F. Manly, J. Grafman, Neural correlates of automatic beliefs about gender and race, Human Brain Mapping 28 (2007) 915–930. [19] Q. Luo, M. Nakic, T. Wheatley, R. Richell, A. Martin, R.J.R. Blair, The neural basis of implicit moral attitude—an IAT study using event-related fMRI, NeuroImage 30 (2006) 1449–1457. [20] M.A. Nitsche, A. Roth, M.F. Kuo, A.K. Fischer, D. Liebetanz, N. Lang, F. Tergau, W. Paulus, Timing-dependent modulation of associative plasticity by general network excitability in the human motor cortex, Neuroscience 27 (2007) 3807–3812.
86
T.E. Gladwin et al. / Neuroscience Letters 517 (2012) 82–86
[21] B.A. Nosek, A.G. Greenwald, M.R. Banaji, Understanding and using the Implicit Association Test: II. Method variables and construct validity, Personality and Social Psychology Bulletin 31 (2005) 166–180. [22] S.H. Ohn, C.I. Park, W.K. Yoo, M.H. Ko, K.P. Choi, G.M. Kim, Y.T. Lee, Y.H. Lee, Timedependent effect of transcranial direct current stimulation on the enhancement of working memory, Neuroreport 19 (2008) 43–47. [23] A.M. Owen, N.J. Herrod, D.K. Menon, J.C. Clark, S.P. Downey, T.A. Carpenter, P.S. Minhas, F.E. Turkheimer, E.J. Williams, T.W. Robbins, B.J. Sahakian, M. Petrides, J.D. Pickard, Redefining the functional organization of working memory processes within human lateral prefrontal cortex, European Journal of Neuroscience 11 (1999) 567–574. [24] M. Petrides, The role of the mid-dorsolateral prefrontal cortex in working memory, Experimental Brain Research 133 (2000) 44–54. [25] A. Roefs, J. Huijding, F.T.Y. Smulders, C.M. MacLeod, P.J. de Jong, R.W. Wiers, A.T. Jansen, Implicit measures of association in psychopathology research, Psychological Bulletin 137 (2011) 149–193.
[26] J.W. Sherman, B. Gawronski, K. Gonsalkorale, K. Hugenberg, T.J. Allen, C.J. Groom, The self-regulation of automatic associations and behavioral impulses, Psychological Review 115 (2008) 314–335. [27] E.E. Smith, J. Jonides, Neuroimaging analyses of human working memory, Proceedings of the National Academy of Sciences of the United States of America 95 (1998) 12061–12068. [28] R. Sparing, M. Dafotakis, I.G. Meister, N. Thirugnanasambandam, G.R. Fink, Enhancing language performance with non-invasive brain stimulation—a transcranial direct current stimulation study in healthy humans, Neuropsychologia 46 (2008) 261–268. [29] N. Unsworth, R.W. Engle, On the division of short-term and working memory: an examination of simple and complex span and their relation to higher order abilities, Psychological Bulletin 133 (2007) 1038–1066. [30] X. Zhang, L. Ma, S. Li, Y. Wang, X. Weng, L. Wang, A mismatch process in brief delayed matching-to-sample task: an fMRI study, Experimental Brain Research 186 (2008) 335–341.