The Power of Sham
Electrical stimulation of the dorsolateral prefrontal cortex
improves working memory for real and placebo
stimulation
Iris de Vries 10367675 Bachelor thesis
Faculty of Social and Behavioral Sciences, Department of Psychology Thesis Advisors: Dr. I. G. Sligte, A. Laufer MSc
Abstract
Even though we experience the visual world around us as highly detailed, our capacity to maintain visual information in mind is highly limited. Yet, we constantly rely on this so-called visual working memory, as it is essential for higher cognitive functions and fluid intelligence. Based on previous research, we hypothesized that visual working memory relies on brain activity in the right dorsolateral prefrontal cortex (DLPFC). In this experiment, we investigated whether working memory capacity could be enhanced by stimulating the right DLPFC with weak electrical current. In different stimulation sessions, participants either received sham/placebo or anodal stimulation, and the order was counterbalanced over participants. Importantly, both test leader and participant were blind to whether sham or real stimulation was applied. While we predicted that performance would only increase during the real stimulation session, we observed an improvement in performance regardless of whether real or sham stimulation was used. This strongly indicates that the improvement in working memory capacity was caused by a placebo effect, while the actual current had no effect on working memory capacity. This finding emphasizes the importance of double-blind placebo-controlled experiments in brain stimulation studies. For now, we conclude that electrical stimulation of the right DLPFC does not enhance visual working capacity. Introduction
Humans face a complex environment everyday. Different sensory organs and processing mechanisms process different aspects of the external world. An important part of the external incoming information is visual information. It enters via the eyes and it is briefly maintained by visual working memory after which it can be stored in long term memory (Miller, Erickson, & Desimone, 1996; Vogel, & Machizawa, 2004). Visual working memory is helpful in many situations, for example to compare scenes with each other to detect changes that indicate danger. Imagine you are on vacation in Australia and notice a dangerous spider on the floor of your apartment. Because you have looked at it for quite some time and it did not move, you conclude it must be dead. You look away to press the button of the vacuum cleaner to remove the spider, but when you turn around you notice something has changed. The spider seems to have moved a little. If it was not for the image that was stored in visual working memory, you would not have noticed the movement of the spider, which indicated it was still alive and very
dangerous. This example illustrates the fact that we constantly rely on visual working memory, as it is essential for higher cognitive functions and fluid intelligence (Engle, Tuholski, Laughlin, & Conway, 1999). Generally, people are fairly good at using their visual working memory. Therefore, it seems like the external world is rich and continuous, like an exact representation of reality.
However, the change detection task, a typical task to measure visual working memory performance (Luck & Vogel, 2013), indicates that this is not the case. In a visual change detection task (see Figure 1), a couple of items are presented in a memory array, and people are asked to remember as much items as possible during a blank retention interval. After that, a probe item appears and the subjects have to report whether the probe item is identical to the item in the memory array or whether it has changed. This is usually not a problem when it concerns four items or less. It does, however, become difficult when the number of items within a set increases. The common explanation for this phenomenon is that the capacity of visual working memory is limited to about three to four separate items. (Todd & Marois, 2004; Rensink, 2002; Simons, & Levin, 1997). Despite the fact that people are managing to interact with the (visual) environment efficiently, the capacity of visual working memory is extremely limited.
Enhancing visual working memory capacity seems an important issue due to its relevance in daily life. Recently, brain stimulation has gained popularity and several types of brain stimulation, such as transcranial magnetic stimulation (TMS) and transcranial direct current stimulation (tDCS), have been shown to be very effective in altering cognitive functioning (Coffman, Clark, & Parasuraman, 2014; Oliveri et al., 2001) Therefore, in order to enhance visual working memory capacity, the use of brain stimulation is an obvious choice.
Brain stimulation requires knowledge about the exact brain areas involved in visual working memory. Visual working memory capacity seems to depend on neural activity in a widespread network of brain regions. This is supported by functional neuroimaging studies in humans (Courtney et al., 1997; Smith & Jonides, 1999). Many areas are found to be part of this network, for example, Todd and Marois (2004) showed that the posterior parietal cortex activity neurally reflects the capacity limit. Areas in the dorsolateral and ventrolateral prefrontal cortex (DLPFC, VLPFC), the superior, inferior parietal lobule (SPL, IPL) and the pre-SMA (Linden et al, 2003; Pessoa, Gutierrez, Bandettini, & Ungerleider, 2002) also seem to part of the network. It has been shown
that the DLPFC remains active during the delays of visual working memory tasks (Curtis & D’Esposito, 2003) and that lesions in this region impair performance on working memory tasks. These last findings provide evidence for the special status of the DLPFC concerning visual working memory capacity.
The question that arises is whether there is a causal relation between DLPFC activity and working memory capacity. If that is the case, stimulating a specific brain area might in fact influence visual working memory capacity. Interestingly, transcranial magnetic stimulation (TMS) studies showed that right DLPFC stimulation during a working memory task negatively influences working memory capacity (Oliveri et al., 2001; Turatto, Sandrini, & Miniussi, 2004; Sligte, Wokke, Tesselaar, Scholte, & Lamme, 2011). This indicates that there is a causal relation between right DLPFC activity and working memory and that it is possible to influence visual working memory by stimulating this region.
A virtual lesion caused by TMS can thus decrease visual working memory capacity. We therefore hypothesize that visual working memory capacity depends on DLPFC excitability. Stimulating the right DLPFC in a way that excites the underlying cortex may increase the capacity. In contrast to the virtual lesion that can be caused by TMS, anodal transcranial direct current stimulation (tDCS) is a method that is able to raise the excitability of the underlying cortex. It uses a constant, low current that is delivered to the brain via electrodes placed directly on the scalp. Positive polarization (anodal) induces excitation of the underlying cortex area, while negative polarization (cathodal) induces inhibition. Recent tDCS studies have already shown improvement of attention, learning, and general working memory capacities (Coffman, Clark, & Parasuraman, 2014; Fregini et al., 2005). Because of promising earlier research on other cognitive functions and the excitability effects on the cortex of anodal tDCS stimulation, tDCS applied to the right DLPFC seems like an excellent method to enhance visual working memory capacity.
In this experiment, anodal tDCS is compared to sham (placebo) stimulation in order to investigate the effects of tDCS to the DLPFC on visual working memory capacity. We hypothesized that visual working memory capacity depends on DLPFC excitability and we expect anodal tDCS to increase performance on the change detection task. We do not expect an increased performance during sham stimulation. A schematic overview of the predictions is shown in Figure 2a.
Figure 1. The change detection task schematically displayed with the duration of each part between brackets. The orientation of the probe rectangle shown in this figure did not change. Methods
Participants
Eighteen subjects participated in the experiment. Two subjects dropped out of the experiment due to planning difficulties. This resulted in 16 participants completing the experiment. Five of them were males and 11 of them were females with the mean age of 20 years (SD = 1.79). Participants could only participate when they did not meet any exclusion criteria. The main criteria were: a personal or family history of epilepsy (because of the risk of an epileptic event caused by tDCS), left-handedness (because of lateralization differences between left-handed and right-handed people), a history of migraine or heavy headaches and dizziness, and neurological disorders (tDCS might worsen the symptoms), and bad vision or color blindness. No financial reward was given, but participants could be rewarded with a psychology research credit per hour of participation. Subjects gave their written informed consent before participation in this study, which was approved by the local ethics committee.
Materials
The task was programmed in Presentation and the stimuli were presented on a 24-inch LCD monitor with a resolution of 1920 by 1080 pixels. Participants were seated 57 cm of the screen. The background of the screen maintained light grey (224(R) 224(G) 224(B)) throughout the whole experiment.
The task in this experiment was based on a classic change detection task (Luck & Vogel, 2013). A schematic overview of the task is show in Figure 1. A fixation cross (20 x 4 pixels) was shown in the middle of the screen for 1000 ms after onset of a trial. After that, the cross turned green indicating an upcoming memory array 500 ms later. The
array consisted of eight black rectangles (25 x 100 pixels) placed on the screen in a way that a circle was formed around the fixation cross. The radius was 15% of the total screen size. The first rectangle was placed at position 0 on the unit circle and the other rectangles were placed along the circle in steps of 1/8 π. The possible angles of the rectangles were: 0◦, 45◦, 90◦, and 135◦. Rectangles differed in their orientation in a way that it was not possible for one orientation to appear more than three times. The memory array was displayed for 250 ms followed by the black fixation cross for 1000 ms. After that rectangle positions were randomly chosen to reappear as probe. On 50% of the trials, the probe changed to an orthogonal orientation (i.e. changed 90◦). Next the probe was visible for 3000 ms during which participants had to decide whether they thought the probe had changed orientation. A press on the z key indicated no change and the m key indicated change. To prevent confusion, the z key was marked with a red sticker and the m key was marked with a green sticker. A short sound was used to inform participants about a correct performance. When not pressing z or m, the trial was saved as a missed trial.
In total, there were three sessions containing three blocks of 20 minutes. Each block consisted of 160 trials. One trial took 5750 ms in total. Independent of the input of the participants, each trial took the complete 5750 ms before starting a new trial.
Sham Condition Versus Experimental Condition
In order to perform this experiment double blind and placebo controlled, a sham condition was added to the research design. The tDCS device was programmed in a way that different codes stood for tDCS or sham stimulation. Neither the experimenter nor the participant was aware of the kind of stimulation.
An anodal electrode of 9 cm2 was placed on area F4, above the right DLPFC. A cathodal electrode of 35 cm2 was placed on the contralateral supraorbital area, the left part of the forehead. These locations were determined by using an EEG cap with the 10-20 system. Conductivity was enhanced by a salt solution (two tea spoons/L). When applying anodal tDCS, direct current of 1 mA was sent through the scalp for 20 minutes during the second block of the session. Sham stimulation was achieved by a normal ramp-up period of 60s. Next, the current was turned off until the end of the experiment when it was turned back on again, 60s before the ending of the second block. During these ramp-up and ramp-down periods, participants experience the same mild
uncomfortable feeling on the location of the electrodes as with tDCS. This is similar to the experience of the real tDCS because of desensitization during tDCS stimulation. Procedure
The recruited participants were sent an information brochure to inform them about the experiment beforehand. They were called to make three appointments with at least 72 hours between sessions. The first session was a training session of one hour. The next two sessions were experimental sessions. In the training session, participants were first provided with information about the experiment and gave their written informed consent. The rest of the hour, participants had the chance to get familiar with tDCS and the change detection task itself. The next two sessions, participants underwent anodal tDCS or sham stimulation. In order to control for experimenter effects, the experiment was double blind. During the first experimental session, a randomly chosen code that corresponded with real tDCS or sham stimulation determined which one was performed first. During the second experimental session a code for the opposite setting was used.
Each experimental session took approximately 1 hour. The experiment consisted of three blocks (20 minutes each) with 160 change detection trials per block with a break of 90s every 5 minutes. Between blocks, there was an additional break of 90s. During the first block, participants performed the change detection task without stimulation. During the second block, tDCS or sham stimulation was delivered. Stimulation was turned off during the third block. After each experimental session, the participant could choose to be brought home by one of the experimenters. After the last experimental session, we provided the participants with a debriefing. There was an opportunity to ask questions afterwards.
Analysis
The data were analyzed with MATLAB software version 8.4.0.150421 (R2014b). Mean performance was calculated for each participant on each block of each session. For participant 4, the last block of the first (training) session was missing because of a power failure. The missing block and all other missed trials and responses in the data were replaced by the build-in function Not-a-Number (NaN). In this way, these data were excluded from analysis.
The statistical analysis was done in IBM SPSS 23. To test for sphericity, Mauchly’s Test of Sphericity was performed. It indicated that the assumption of sphericity had not been violated, χ2(2) = 0.065, p = .968, χ2(2) = 1.665, p = .435. Within each session, a check for outliers was done. If a participants’ performance differed more than three standard deviations from the mean, the participant was excluded from the sample. No participants were deleted.
A repeated measures ANOVA was performed to look for interaction and main effects. The ANOVA was performed on the mean scores per participant per session per block; with time over one session (three levels) and stimulation (two levels) as within-subject variables.
Results
In this study, an attempt was made to enhance visual working memory capacity. We did this by applying tDCS or sham to the right DLPFC of 16 participants while they were performing a change detection task. In Table 1, the mean scores and standard deviation of the change detection tasks for the different blocks and sessions are displayed. Based on the literature, we predicted equal performance across blocks for the sham condition and an increase in performance on block 2 (during tDCS) for the anodal tDCS condition (see Figure 2a). However, this predicted interaction was not found (F(2, 30) = 0.534 , p = .591). In both conditions, we observed an improvement in performance regardless of whether anodal tDCS or sham stimulation was used (F(2, 30) = 5.52, p = .009). This indicates a placebo effect. tDCS does not seem to differ from sham stimulation as a whole (F(1, 15) = 0.782, p = .390). See Figure 2b for the graphic display of the actual results. When comparing the combined mean scores of tDCS and sham of block 1 and block 2, a 4% overall increase in performance was found. This increase in performance during the second block can be attributed to the placebo effect.
Figure 2a. The expected performance Figure 2b. The actual performance on the on the change detection task in the the change detection task in the tDCS tDCS and the sham condition and the sham condition.
with training as block 1, tDCS with training as block 1, tDCS as block 2, and sham as block 3. as block 2, and sham as block 3.
If these results could be attributed to learning effects, you would expect differences between the last block of the training session and the first block of both experimental sessions. This is because participants would still be getting better at the task at the two experimental sessions after the completing of the training session. If it is not an effect of learning, participants reached their maximal level of performance after the third block of the training session. If that is the case, there would be no differences between the third block of the training session and the first block of the experimental sessions. Performance on the last block of the training session did not differ from the first block of the sham session (t(14) = -0.28, p = .978) or the first block of the tDCS session (t(14) = -0.160, p = .875). These non-significant results indicate that the improvements could not be attributed to learning effect.
Table 1. The mean scores and the SD (between brackets) of the change detection tasks for the different blocks and sessions are displayed.
We exploratory looked at whether reaction time differed between anodal tDCS and sham stimulation, because we hypothesized that the effect of tDCS could be reflected in reaction times instead of percentage correct on the change detection task. However, reaction times did not differ between the anodal tDCS and the sham stimulation (t(15) = .672, p= .512).
Discussion
Humans face a complex environment everyday. Despite the importance of visual working memory in daily life and the fact that people are managing to interact with the (visual) environment efficiently, the capacity of visual working memory is extremely limited. Therefore, enhancing visual working memory is interesting. Research indicates a special status of the DLPFC concerning visual working memory and TMS research even indicates a causal relation between DLPFC activity and working memory capacity. We hypothesized that visual working memory capacity depends on the excitability of the right DLPFC. Because of promising earlier research on other cognitive functions and the excitability effects on the cortex of anodal tDCS stimulation, tDCS applied to the right DLPFC seemed like an excellent method to enhance visual working memory capacity. We expected that performance would increase during tDCS stimulation and no differences were expected during the sham stimulation.
The following two major findings and implications are going to be discussed in this section: (1) in contrast to the expectation that participants would perform better under tDCS stimulation, it was found that this was not the case and that tDCS did not enhance visual working memory capacity, and (2) in both conditions, we observed an improvement in performance regardless of whether anodal tDCS or sham stimulation was used. This indicates a placebo effect.
In contrast to the expectation that participants would perform better when tDCS was applied, it was found that this was not the case and that tDCS did not enhance visual working memory capacity. The choice of using tDCS on the DLPFC was mainly based on research concerning TMS. In these studies, creating a ‘virtual lesion’ by applying TMS to the DLPFC induced a decrease in visual working memory performance (Oliveri et al., 2001; Turatto, Sandrini, & Miniussi, 2004; Sligte et al., 2011). We expected that the capacity would increase when trying to achieve the opposite of a ‘virtual lesion’ by using tDCS. These contradicting findings might be due to a couple of possible factors. Firstly,
the sample size was very limited. Due to the small effects of tDCS, a large sample is required to detect an effect. Further, there might be an unchangeable limit of visual working memory that cannot be exceeded in any way. The fact that the capacity can be decreased does not necessarily mean that it can also be increased. Another issue that needs to be considered is the fact that the task was quite monotonous, long-winded, difficult, and required constant fast reactions. Much attention was needed to answer trials correctly. It may be that the task did not measure visual working memory, but did in fact measure attention. Attention processes might not depend on excitability of the DLPFC. Lastly, even though the choice of stimulating the DLPFC is supported by a large body of research, it might not be enough to stimulate the DLPFC in isolation.
Interestingly, it was found that performance improved in both conditions regardless of whether anodal tDCS or sham stimulation was used, indicating a placebo effect. This placebo effect might be due to the mere presence of the tDCS apparatus. The idea of the stimulation and the expected effects of this device might have increased the performance. Strange enough, anodal stimulation seems to increase performance less than sham stimulation. However, this is only speculation and needs to be clarified in future research by using an active control (cathodal stimulation as placebo condition).
An important lesson that can be learned from this study is that it is necessary to control for placebo effects in brain stimulation studies too. Because of the seemingly ‘hard’ method of brain stimulation, it is often neglected, forgotten or simply thought of as redundant to use a placebo condition in brain stimulation research. The article of Horvath, Forte, and Carter (2015) includes a substantial body of brain stimulation research. Throughout the whole article, it becomes clear how many of the included research lacks a proper design. Often, sham conditions are not included and the experiment is not set up double blind. For example, the authors stated that “the lack of sham controlled, double-blind (…) TMS studies are somewhat alarming”. In their study they tested a large body of brain stimulation research and came to the conclusion that only 18 of the 106 TMS studies included in the analysis used a sham condition. Additionally, only 7 of the 106 studies used a double blind design. In addition to the previous article, another study revealed that only 87.5% of the tDCS studies examining foundational claims, on which the modern tDCS field is built, used a proper control condition (Horvath, Carter, & Forte, 2014).
In the future, stricter guidelines should be available on the use of sham or active control (cathodal stimulation) conditions. Further, brain stimulation studies always need a double-blind design. Studies without a double blind and controlled set-up contribute to the framework of literature, possibly causing a distorted view of reality. This might result in hypotheses based on that view. To prevent the confusion of placebo effects with actual effects, it is critical to investigate placebo effects and the use of a double blind setup in brain stimulation studies.
This study showed that the use of tDCS to enhance visual working memory is not straightforward. Effects of DLPFC activity on visual working memory capacity were not found, but a placebo effect was. Further research on the effects of tDCS of the right DLPFC on visual working memory should at least focus on a bigger sample size. More importantly, it should be double blind and placebo controlled. No matter how straightforward the instrument and literature seems, never underestimate the power of suggestion. For now, there seems to be no proof that anodal tDCS enhances visual working memory when stimulating the DLPFC. Replication needs to be done to clarify why we did not find an effect of tDCS on visual working memory and to determine the contribution of the placebo effects in brain stimulation.
References
Coffman, B. A., Clark, V. P., & Parasuraman, R. (2014). Battery powered thought:
enhancement of attention, learning, and memory in healthy adults using transcranial direct current stimulation. Neuroimage, 85, 895-908.
Courtney, S. M., Ungerleider, L. G., Keil, K., & Haxby, J. V. (1997). Transient and sustained activity in a distributed neural system for human working memory. Nature,
386(6625), 608-611.
Cowan, N., Elliott, E. M., Saults, J. S., Morey, C. C., Mattox, S., Hismjatullina, A., & Conway, A. R. A. (2005). On the capacity of attention: Its estimation and its role in working memory and cognitive aptitudes. Cognitive Psychology, 51, 42-100.
Curtis, C. E., & D'Esposito, M. (2003). Persistent activity in the prefrontal cortex during working memory. Trends in cognitive sciences, 7(9), 415-423.
Engle, R. W., Tuholski, S. W., Laughlin, J. E., & Conway, A. R. (1999). Working memory, short-term memory, and general fluid intelligence: a latent-variable approach.
Fregni, F., Boggio, P. S., Nitsche, M., Bermpohl, F., Antal, A., Feredoes, E., ... & Pascual Leone, A. (2005). Anodal transcranial direct current stimulation of prefrontal cortex enhances working memory. Experimental brain research, 166(1), 23-30. Horvath, J. C., Carter, O., & Forte, J. D. (2014). Transcranial direct current stimulation:
five important issues we aren't discussing (but probably should be). Frontiers in
systems neuroscience, 8.
Horvath, J. C., Forte, J. D., & Carter, O. (2015). Evidence that transcranial direct current stimulation (tDCS) generates little-to-no reliable neurophysiologic effect beyond MEP amplitude modulation in healthy human subjects: A systematic review.
Neuropsychologia, 66, 213-236.
Linden, D. E., Bittner, R. A., Muckli, L., Waltz, J. A., Kriegeskorte, N., Goebel, R., ... & Munk, M. H. (2003). Cortical capacity constraints for visual working memory:
dissociation of fMRI load effects in a fronto-parietal network. Neuroimage, 20(3), 1518-1530.
Luck, S. J., & Vogel, E. K. (2013). Visual working memory capacity: from psychophysics and neurobiology to individual differences. Trends in cognitive sciences, 17(8), 391-400.
Miller, E. K., Erickson, C. A., & Desimone, R. (1996). Neural mechanisms of visual working memory in prefrontal cortex of the macaque. The Journal of Neuroscience, 16(16), 5154-5167.
Oliveri, M., Turriziani, P., Carlesimo, G. A., Koch, G., Tomaiuolo, F., Panella, M., & Caltagirone, C. (2001). Parieto-frontal interactions in visual-object and visual spatial working memory: evidence from transcranial magnetic stimulation.
Cerebral Cortex, 11(7), 606-618.
Pessoa, L., Gutierrez, E., Bandettini, P. A., & Ungerleider, L. G. (2002). Neural correlates of visual working memory: fMRI amplitude predicts task performance. Neuron,
35(5), 975-987.
Rensink, R. A. (2002). Change detection. Annual review of psychology, 53(1), 245 277.
Simons, D. J., & Levin, D. T. (1997). Change blindness. Trends in cognitive sciences, 1(7), 261-267.
stimulation of the dorsolateral prefrontal cortex dissociates fragile visual short-term memory from visual working memory. Neuropsychologia, 49(6), 1578-1588.
Smith, E. E., & Jonides, J. (1999). Storage and executive processes in the frontal lobes.
Science, 283(5408), 1657-1661.
Todd, J. J., & Marois, R. (2004). Capacity limit of visual short-term memory in human posterior parietal cortex. Nature, 428(6984), 751-754.
Turatto, M., Sandrini, M., & Miniussi, C. (2004). The role of the right dorsolateral prefrontal cortex in visual change awareness. Neuroreport, 15(16), 2549-2552. Vogel, E. K., & Machizawa, M. G. (2004). Neural activity predicts individual differences in
visual working memory capacity. Nature, 428(6984), 748-751.
Wirth, M., Rahman, R. A., Kuenecke, J., Koenig, T., Horn, H., Sommer, W., & Dierks, T. (2011). Effects of transcranial direct current stimulation (tDCS) on behaviour and electrophysiology of language production. Neuropsychologia, 49(14), 3989-3998.
