Measuring Thought-Control Failure: Sensory Mechanisms and Individual Differences

Abstract

The ability to control one’s thoughts is crucial for attention, focus, ideation, and mental well-being. Although there is a long history of research into thought control, the inherent subjectivity of thoughts has made objective examination, and thus mechanistic understanding, difficult. Here, we report a novel method to objectively investigate thought-control success and failure by measuring the sensory strength of visual thoughts using binocular rivalry, a perceptual illusion. Across five experiments (N = 67), we found that thought-control failure may occur because of the involuntary and antithetical formation of nonreportable sensory representations during attempts at thought suppression but not during thought substitution. Notably, thought control was worse in individuals with high levels of anxiety and schizotypy but more successful in mindful individuals. Overall, our study offers insight into the underlying mechanisms of thought control and suggests that individual differences play an important role in the ability to control thoughts.

Keywords

thought suppression thought substitution binocular rivalry mental imagery mindfulness open data

The ability to control our own thoughts is a crucial mental capacity important for attention, focus, and ideation (Purdon & Clark, 2002; Wegner, 1989). Given that up to 80% of the general population experience some form of unwanted intrusive thoughts (Rachman & de Silva, 1978), this ability is also an important determinant of mental well-being. Indeed, failure to control thoughts is linked to various mental disorders, including anxiety (Freeston, Dugas, & Ladouceur, 1996), obsessive-compulsive disorder (Rachman & de Silva, 1978), and schizophrenia (Chapman & Chapman, 1973).

Since the work of Freud, the idea of voluntarily repressing a thought, or thought suppression, has become an active phrase in the general populace and attracted much attention. Attempting to control thoughts by thought suppression (not thinking about a given thought) is often difficult, and subjective reports suggest that suppression mostly fails, paradoxically leading to heightened preoccupations with these thoughts (Wegner, 1989). For example, individuals instructed to suppress the thought of a white bear are often unable to do so and report intrusions of the thought inadvertently arising a short time later (Wegner, Schneider, Carter, & White, 1987). Thought-suppression failure and its consequences have been studied in a wide variety of domains including organizational behavior (Kulik, Perry, & Bourhis, 2000), immune function (Petrie, Booth, & Pennebaker, 1998), and psychopathology (Wegner, 1989).

The inability to suppress one’s thoughts is attributed to the interaction between two practical, but simultaneously conflicting, mental processes (Wegner, 1994). The first involves a conscious process that attempts to achieve a state of mind free from the to-be-suppressed thought. The second is an unconscious monitoring process that checks for unwanted instances of the to-be-suppressed thought. It is proposed that thoughts unwillingly enter consciousness because of the ironic opposition between these two processes, ultimately leading to thought-control failure. Neuroimaging evidence has generally supported this view (Giuliano & Wicha, 2010; Mitchell et al., 2007).

While thought suppression is the most commonly studied thought-control strategy, other forms of thought control have been examined and appear to be more effective. For example, thought substitution, in which a suppressed thought is replaced by a substitute thought, has shown evidence for reduced thought-intrusion frequency and thought-control failure (Lin & Wicker, 2007; Wegner, 1989). More recently, mindfulness has been shown to be an effective form of mental control (Brown & Ryan, 2003).

Despite research into these various forms of thought control, the inherent subjectivity of thoughts has made the objective examination of thought control difficult. For the most part, researchers have previously used subjective self-reports to examine thought control (e.g., Lavy & van den Hout, 2009; Lin & Wicker, 2007; Wegner et al., 1987). Although self-report measures have provided valuable insight, these measures may be prone to bias, experimenter demand, or social-desirability effects (Stone, 2000), making objective investigation of underlying mechanisms difficult.

To shed light on the underlying mechanisms of thought control and its failure, we devised a novel method to objectively study the control of visual thoughts. We used the visual illusion binocular rivalry, which arises when an image is presented to one eye while a different image is simultaneously presented to the other eye, inducing bistable perceptual alternations between the two images (Alais & Blake, 2005). Importantly, binocular rivalry has been used to empirically investigate visual mental imagery (Chang, Lewis, & Pearson, 2013; Keogh & Pearson, 2011, 2017, 2018; Pearson, 2014; Pearson, Clifford, & Tong, 2008; Pearson, Naselaris, Holmes, & Kosslyn, 2015) and forms of synesthesia (Chiou, Rich, Rogers, & Pearson, 2018). Consequently, we used visual thoughts, specifically visual mental imagery, and a devised thought-control index to objectively investigate the nature and sensory mechanisms of thought control. We then examined individual differences in thought control across four psychological traits (anxiety, obsessive-compulsive tendencies, schizotypy, and mindfulness).

Results showed that the sensory representation of suppressed visual thoughts was almost as strong as that for intentionally imagined thoughts, indicating the failure to control the sensory trace of thoughts using thought suppression. Strikingly, visual thoughts reported as successfully suppressed (i.e., without overt visual experience) formed a measurable sensory trace, suggesting the possibility of emergent nonconscious sensory representations during attempts at thought control. Thought substitution was more, although not completely, effective in controlling the sensory trace of visual thoughts. Individual differences showed that individuals high in trait mindfulness exhibited greater levels of thought control, whereas individuals high in trait anxiety or schizotypy exhibited lower levels of thought control. Interestingly, there was no relationship with obsessive-compulsive tendencies. Overall, our data show that thought-control failure may be linked to the formation of sensory representations during attempts at thought control and suggest that individual differences across psychological traits play an important role in the ability to control thoughts.

Method

Participants

Ten participants (3 female; age: M = 25.60 years, SD = 4.62) completed the thought-suppression, luminance, location, and thought-substitution experiments. The target sample size was determined by a power analysis of pilot experiment data and also reflected the number of participants used in past research on binocular rivalry and visual mental imagery (e.g., Chang et al., 2013; Keogh & Pearson, 2017; Pearson et al., 2008; Sherwood & Pearson, 2010). All participants were the same across these four experiments. However, 2 participants who were unable to complete the location experiment were replaced by 2 age- and gender-matched participants.

The individual-differences experiment consisted of a total of 67 participants (35 female; age: M = 21.88 years, SD = 5.03), which included the data from the 10 participants in the main experiment. The target sample size was determined by a power analysis of pilot experiment data and also reflected the number of participants used in previous correlational research on individual differences in mental-imagery strength (e.g., Kosslyn, Brunn, Cave, & Wallach, 1984). Written consent was obtained from all participants, and the experiment was approved by the ethics committee of the University of New South Wales.

Materials and procedure

Thought-suppression experiment

All experiments followed the design of the thought-suppression experiment (see Fig. 1a). The task was carried out on a Windows 7 PC running Psychophysics Toolbox (Version 3; Brainard, 1997) in MATLAB (The MathWorks, Natick, MA) on an 85-Hz Dell Trinitron P1130 CRT monitor with 1,280 × 1,024 resolution. Red-green colored filters, one over each eye, were worn by participants to view the binocular-rivalry stimulus: a red-green Gabor patch. A chin rest positioned each participant’s head 57 cm from the screen. Before the experiment, each participant completed an eye-dominance test (as reported by Pearson et al., 2008; Sherwood & Pearson, 2010) to determine the stimulus intensity at which the red and green grating stimuli were equally dominant for each participant.

Fig. 1.

Measuring thought control with binocular rivalry—time course and conditions of the thought-suppression, luminance, location, and thought-substitution experiments. In the thought-suppression experiment (a), each participant was randomly shown one of three red (“red apple,” “red chili,” “red tomato”) or green (“green broccoli,” “green cucumber,” “green lime”) written object cues and instructed to either “imagine” or suppress (“avoid imagining”) the thought of that object for 7 s. In suppression trials, participants pressed an assigned button to indicate a thought intrusion—a failed suppression attempt—during the thought-control period. Participants were then briefly shown a red-green binocular-rivalry stimulus before reporting binocular-rivalry dominance. The procedure for the other three experiments shown here was the same, except as follows. In the thought-substitution experiment (b), participants instead imagined a white object (“white cloud,” “white sheep,” “white cauliflower”). In the luminance experiment (c), participants viewed changing background luminance during the thought-control period. In the location experiment (d), participants viewed the rivalry stimulus in one of four locations (top, bottom, left, right) around the screen during binocular rivalry (only the top location is shown here). Vividness, effort, memory control, and decision-bias control questions are not shown.

The experiment consisted of a minimum of three blocks of 42 trials (minimum total of 126 trials) over a 1-hr session in a dark room. If time permitted, additional blocks were performed. Participants focused on a fixation point at the center of the screen for each trial, and a black background was present throughout the experiment.

Each trial comprised four stages: (a) object presentation, (b) thought instruction, (c) thought-control period, and (d) binocular rivalry. In the object-presentation stage, words describing one of six object stimuli (three green objects—“green broccoli,” “green lime,” “green cucumber”; three red objects—“red chili,” “red apple,” “red tomato”) were pseudorandomly presented on the screen as a cue. Next, in the thought-instruction stage, the words “imagine” or “avoid imagining” appeared on the screen. In the imagine condition, participants actively imagined the presented object. In the suppression or avoid-imagining condition, participants attempted to suppress any thoughts of that object. The two conditions were presented in random order to remove order effects and were intermixed rather than presented in separate blocks to increase participant engagement in the task. Then, in the thought-control period, participants either imagined or suppressed the thought of the object stimulus for 7 s, depending on the thought instruction provided in the previous stage. If participants failed at thought suppression, they were instructed to press the space bar on the keyboard. Lastly, in the binocular-rivalry stage, a red-green Gabor patch appeared in the center of the screen (red horizontal: luminance = 13.2 cdm², Commission Internationale de l’Éclairage, or CIE color space: x = 0.360, y = 0.358; green vertical: luminance = 10.9 cdm², CIE color space: x = 0.263, y = 0.568), and participants perceived binocular rivalry between the colors red and green (0.75 s). Participants indicated, by pressing assigned keys, whether they perceived (a) color dominance in red, (b) color dominance in green, or (c) an equal mix of both colors (50% red, 50% green).

Before the experiment, participants were explicitly told that during the suppression trials, they were not to use another visual object to substitute or distract them from the object they were trying to suppress. Furthermore, it was asserted that there was no right or wrong response and they were to respond as truthfully as possible.

Each block of 42 trials contained 36 experimental trials. Within the 36 experimental trials, there were an equal number of imagery and suppression trials (18 imagine, 18 avoid imagining), and trials of the object stimuli were divided equally among the 6 objects (6 “red apple,” 6 “green cucumber,” etc.; total of 18 red objects and 18 green objects). The remaining 6 trials consisted of memory-control trials. At the end of these trials, a memory-control question required participants to correctly identify the cued object presented to them during the object-presentation stage (i.e., the object they were instructed to imagine or suppress) from a list of 3 object stimuli. These memory-control questions were presented in place of the binocular-rivalry stimulus and ensured that participants were paying attention to the stimuli to be imagined or suppressed. The order in which the 42 trials were presented in each block were randomized to eliminate order effects.

At the conclusion of the experiment, participants were asked to provide a verbal report on their performance, including whether they found thought control to be easy or difficult, as well as the effectiveness of the strategies used during thought control.

Luminance experiment

In the luminance experiment (see Fig. 1c), we examined whether the sensory trace of suppressed thoughts could be disrupted by bright uniform background luminance. The procedure for the luminance experiment was identical to that in the main thought-suppression experiment except for one modification. During the thought-control period, the background luminance smoothly transitioned from black to yellow (luminance = 10.8 cdm², CIE color space: x = 0.375, y = 0.546) over 2 s, remained yellow for 3 s, and then smoothly transitioned back to black (total of 7 s). Yellow was chosen as the background luminance color because it was a mix of the red and green colors of the stimuli and thus would not result in any bias during binocular rivalry.

Location experiment

In the location experiment (see Fig. 1d), we examined whether changing the location of the binocular-rivalry stimulus would affect any priming we observed from suppressed thoughts. The procedure for the location experiment was identical to that of the main thought-suppression experiment except for three modifications. First, the binocular-rivalry stimulus was presented at one of four random locations around the central fixation point (top, bottom, left, right). Second, a random 20% of trials were decision-bias control trials (catch trials), in which a physically fused and equally dominant red-green Gabor patch was presented in place of the standard binocular-rivalry stimulus. Decision-bias control trials were included to test whether participants were veridically reporting binocular-rivalry dominance and to identify any possible nonperceptual bias effects (e.g., demand characteristics and intentional participant misreporting; as reported by, e.g., Keogh & Pearson, 2017; Pearson et al., 2008). These trials consisted of a fake binocular-rivalry stimulus purposely manipulated with a spatial mix of half red and half green to prompt participants to report a mixed response. This Gabor patch was not obviously different from the real binocular-rivalry stimulus, sharing the same size, shape, and color parameters as well as having the same blurred edges and the exact split of red and green varying on each trial to resemble actual rivalry. The third difference between this experiment and the thought-suppression experiment was that two additional questions were included at the end of each trial. A vividness question appeared in all failed suppression trials (when participants reported a thought intrusion) and all imagery trials; this question asked participants to rate how vivid the thought of the object was during the thought-control period. An effort question appeared in all trials and asked participants to rate how effortful it was to imagine or suppress the thought during the thought-control period. Vividness and effort were rated on a scale from 1 (not vivid/effortful) to 4 (very vivid/effortful).

Thought-substitution experiment

In the thought-substitution experiment (see Fig. 1b), we examined the effects of thinking about a substitute-object stimulus, instead of the original-object stimuli, on binocular-rivalry priming. The procedure for the thought-substitution experiment was identical to that in the main experiment except for one modification. In the thought-instruction stage, participants were told to imagine a white object (“white cloud,” “white cauliflower,” or “white sheep”) instead of being told to avoid imagining.

Individual-differences experiment

The individual-differences experiment examined the relationship between thought control and four psychological traits—anxiety, obsessive-compulsive tendencies, schizotypy, and mindfulness. It followed the same procedure as the main thought-suppression experiment (see Fig. 1a). Individual dispositions were measured through four standard questionnaires: the second half of the State-Trait Anxiety Inventory (measuring only trait anxiety and not state anxiety, which was not relevant for this study; Spielberger, 1983), the Obsessive-Compulsive Inventory–Revised (Foa et al., 2002), the Oxford-Liverpool Inventory of Feelings and Experiences (Mason, Claridge, & Jackson, 1995), and the Mindfulness Attention Awareness Scale (Brown & Ryan, 2003).

Statistical analysis

Participants with poor performance (chance or below) on the memory control trials and decision-bias control trials (in the location experiment) were removed from the analysis because they could not be trusted to have been attending to the task or accurately reporting binocular-rivalry dominance. This amounted to 7 participants being removed in the individual-differences experiment because of poor performance on the memory control trials. All participants performed well on the decision-bias control trials, and thus no participants were removed from the analysis in the location experiment. Participants with priming below 40% for imagined or suppressed thoughts were also removed because we were unable to determine the sensory strength of their thought representations: This amounted to 1 participant in the location experiment and 4 participants in the individual-differences experiment.

Priming for each participant was measured by calculating the number of trials in which the dominant rivalry color matched the color of the imagined, suppressed, or substituted stimulus and then dividing this number by the total number of imagery, suppression, or substitution trials, respectively. For the distribution of results shown in Figure 2b, we could not use standard statistics when calculating priming because the number of trials with button presses (i.e., failed thought-control trials) varied across individuals. Consequently, suppression trials from all the participants within each experiment shown in Figure 2b were pooled, and resampling analyses (showing 95% confidence intervals, or CIs) were conducted to calculate priming. Multiple comparisons were controlled by setting the false discovery rate (q) at 0.05 (Benjamini & Hochberg, 1995).

Fig. 2.

Mean percentage of trials that showed priming. Results are shown in (a) for four of the experiments (with the two conditions of the main thought-suppression experiment represented by the two leftmost bars). Asterisks above bars indicate significant differences from chance (p < .05), and asterisks above brackets indicate significant differences between groups (p < .05). Error bars represent standard errors of the mean. Results are shown in (b) for the same four experiments but are broken down by self-reported success or failure during thought suppression. Asterisks indicate significant differences from chance (p < .05). Error bars represent 95% confidence intervals. Multiple comparisons were corrected by controlling the false discovery rate (q) at 0.05.

Results

Thought suppression

We examined whether suppressed thoughts formed a quantifiable sensory trace by measuring how they might affect subsequent binocular-rivalry dominance. We found above-chance binocular-rivalry priming for suppressed thoughts, t(9) = 5.90, p = .001, 95% CI for the mean priming percentage = [58.5, 69.0], d = 1.868, with no difference in priming compared with actively imagined thoughts (p > .05), which also exhibited above-chance priming, t(9) = 5.23, p = .001, 95% CI for the mean priming percentage = [59.0, 72.6], d = 1.652 (see Fig. 2a). In other words, when participants attempted to not think about, for example, a red apple, they were significantly more likely to report red in the subsequent rivalry presentation. This suggests, perhaps surprisingly, that thought suppression was ineffective at preventing the sensory trace of visual thoughts from forming. The probability of self-reported thought-control failure, as indicated by participant button presses, during the main thought-suppression experiment was .29, which was significantly above zero, t(9) = 2.80, p = .041, 95% CI for the probability of thought-control failure = [.056, .524], d = 0.887 (see Fig. S1 in the Supplemental Material available online).

To test whether the priming during attempted thought suppression was of a sensory nature, we used uniform background luminance as a potential sensory perturbation of the formation of a sensory thought representation. Luminance is a visual feature strongly represented in early visual areas (Goodyear & Menon, 1998), including the primary visual cortex (V1; Boynton, Engel, Glover, & Heeger, 1996), and it has been shown that a uniform luminance background can disrupt the formation of intentional mental images (Keogh & Pearson, 2011, 2014, 2017; Pearson et al., 2008; Sherwood & Pearson, 2010). Luminance disruption would thus be expected to result in reduced priming if suppressed visual thoughts are represented in visual areas. Priming for suppressed thoughts in the luminance condition was significantly lower compared with priming with a black background, t(9) = −3.02, p = .024, d = −0.955, but was still above chance, t(9) = 2.71, p = .040, 95% CI for the mean priming percentage = [51.04, 61.57], d = 0.857 (see Fig. 2a), suggesting that priming from suppressed thoughts involved sensory representations in the visual cortex. The probability of self-reported thought-suppression failure in the luminance experiment was .147, which was significantly above zero, t(9) = 2.39, p = .042, 95% CI for the probability of thought-control failure = [.008, .286], d = 0.755, but was lower compared with the main experiment, although not significant (p > .05; see Fig. S1).

Next, we further examined the sensory nature of the priming effect from suppressed thoughts by investigating the local retinotopic nature of the priming bias. If the priming effect was local in visual space, this would suggest that suppressed thought representations are represented in retinotopically organized visual areas. We tested this by changing the spatial location of the binocular-rivalry presentations. Spatial location is a visual feature processed retinotopically (Hubel & Wiesel, 1959), and this retinotopic organization is preserved during mental imagery (Slotnick, Thompson, & Kosslyn, 2005). Consequently, probing the effects of thought suppression at a different retinotopic location (by changing the location of binocular rivalry) would be expected to decrease priming (Chang et al., 2013; Pearson et al., 2008). After the location of binocular rivalry was changed, priming for suppressed thoughts was significantly lower compared with priming in the main thought-suppression experiment, t(8) = −2.47, p = .048, d = −0.802, and was not significantly different from chance (p > .05; see Fig. 2a). The probability of self-reported thought-suppression failure was .46, which was significantly above chance, t(8) = 4.76, p = .006, 95% CI for the probability of thought-control failure = [.235, .678], d = 1.586, and comparable with that of the main thought-suppression experiment (p > .05; see Fig. S1). Combined with the findings from the luminance experiment, these results provide evidence that attempted suppressed thoughts are likely represented in the visual cortex.

In the location experiment, we also explored participant ratings of vividness and the amount of effort used when imagining or suppressing a thought, both rated on a scale from 1 to 4. As shown in Figure S2 in the Supplemental Material, the vividness of imagined thoughts was significantly higher compared with the vividness of suppressed thoughts that arose during failed suppression trials, t(8) = 4.27, p = .003, d = 1.422. No difference was found in the amount of effort expended in imagining or suppressing a thought (p > .05), and there was no significant difference in effort between successful and failed suppression trials (p > .05; see Fig. S2).

To examine individuals’ metacognition of thought control, we compared the subjective reports of thought-suppression success or failure with the level of sensory priming. Priming was greater for trials in which thought suppression was reported as a failure compared with trials reported as successful, although this relationship was not significant (p > .05), suggesting a possible trend toward some metacognition. However, priming was significantly above chance for both successful and failed suppression trials (p = .016, 95% CI for the mean priming percentage = [52, 70]; p < .001, 95% CI for the mean priming percentage = [61, 75]; see Fig. 2b), indicating that the sensory representations of successfully suppressed thoughts were strong enough to drive sensory priming despite the absence of reported awareness. As we expected, there was no difference in priming between successful and failed thought-suppression trials in the luminance and location experiments (p > .05), and neither was significantly different from chance (p > .05; see Fig. 2b).

Thought substitution

We investigated whether imagining a substitute thought instead of the object to be suppressed would be an effective strategy for reducing the perceptual influence of visual thoughts. Previous research has shown that a distraction strategy in which individuals focus on a distractor or substitute thought instead of a target thought can reduce the frequency of reported thought-control failures (Wegner et al., 1987). If this is the case, we might expect priming for substituted thoughts to be lower than that of imagined and suppressed thoughts. Participants were instructed to imagine white objects as a replacement strategy to prevent thoughts about the colored objects from emerging.

Priming in the thought-substitution condition was significantly lower than in both the suppression condition, t(9) = −4.46, p = .005, d = −1.412, and the imagine condition, t(9) = −4.29, p = .005, d = −1.355, of the main thought-suppression experiment and was not significantly above chance (p > .05; see Fig. 2a), suggesting that thought substitution was effective in reducing the sensory trace of visual thoughts. However, failed thought-substitution trials (i.e., when participants reported the occurrence of a thought intrusion) showed above-chance priming (p < .001, 95% CI for the mean priming percentage = [55, 61]; see Fig. 2b) compared with successful thought-substitution trials, which showed no priming effect (p > .05). This suggests that thought substitution was ineffective when the strategy failed but that participants had good metacognition of when such failure occurred. The probability of thought-substitution failure was .132 and was significantly above chance, t(9) = 2.37, p = .042, 95% CI for the probability of thought-control failure = [.006, .258], d = 0.748.

Individual differences in thought control

To investigate individual differences in thought control, we first devised a thought-control index, which measured individuals’ ability to suppress their thoughts. Scores on this index were calculated by finding the difference in priming between imagined thoughts (imagery priming) and suppressed thoughts (suppression priming) for each participant and dividing the result by 100:

(\frac{% imagery priming - % suppression priming}{100}) .

Positive index scores indicate that the sensory strength of imagined thoughts was stronger than suppressed thoughts, suggesting the ability to control visual thoughts. Zero or negative index scores indicate that the sensory strength of suppressed thoughts was as strong as or greater than that of imagined thoughts, suggesting the inability to control visual thoughts. We did not simply use suppression priming as a measure of thought control because this measure would not account for individual differences in mental-imagery strength. Furthermore, by finding the difference between imagery priming and suppression priming (i.e., normalizing by pure imagery priming), we could isolate and examine the nonimagery elements of thought control. Thought-suppression priming was strongly correlated with imagery priming (r = .52, p < .001; see Fig. S3 in the Supplemental Material); this suggested that a good individual measure of thought control was the degree to which suppression priming differed relative to imagery priming.

To confirm that thought-control-index scores were assessing the relative effectiveness of thought control, we applied it to the findings in the thought-suppression, luminance, location, and thought-substitution experiments. As we expected, index scores during thought substitution were significantly higher than index scores during thought suppression, t(9) = 2.32, p = .046, d = 0.732, and were significantly above zero, t(9) = 4.50, p = .006, 95% CI for the mean thought-control-index score = [.061, .184], d = 1.424 (see Fig. S4 in the Supplemental Material), thus confirming the validity of the index.

We used the thought-control index to explore individual differences in the ability to control visual thoughts across four psychological traits (anxiety, obsessive-compulsive tendencies, schizotypy, and mindfulness; see Fig. 3). Index scores were correlated with standardized scores on each of the four questionnaires. Pearson correlations with index scores revealed negative correlations with anxiety (r = −.28, p = .036) and schizotypy (r = −.34, p = .010), indicating that high levels of trait anxiety and schizotypy predicted lower levels of thought control. A positive correlation with mindfulness was found (r = .28, p = .035), indicating that high levels of trait mindfulness predicted greater levels of thought control. No significant correlation was found between index scores and obsessive-compulsive tendencies (r = −.20, p = .143).

Fig. 3.

Scatterplots (with best-fitting regression lines) showing the association between thought-control-index scores and each of four psychological traits (n = 56). Thought-control-index scores were calculated by finding the difference in priming between imagined thoughts and suppressed thoughts for each participant and dividing the result by 100.

Further analyses showed that index scores across all participants followed a normal distribution (minimum = −.127, maximum = .324, SD = .09; see Fig. S5 in the Supplemental Material). The distribution was centered around a mean of 0.045, which was not significantly above zero (p > .05), indicating that participants were, on average, poor at controlling the sensory influence of suppressed visual thoughts.

We verified the validity of our questionnaire data with a correlation matrix for the four traits (see Fig. S6 in the Supplemental Material), which confirmed strong intercorrelations within the negative traits (anxiety, obsessive-compulsive tendencies, schizotypy). In addition, these negative traits were inversely correlated with the positive trait of mindfulness.

Memory control trials

Memory control trials were present in all experiments and monitored whether participants were attending to the prior object-stimuli cues. Aside from 7 participants in the individual-differences experiment who performed poorly (below chance levels) and were removed from the analyses, all participants showed good performance on these trials (M > 85% correct), indicating that participants were properly attending to the task.

Decision-bias control trials

Decision-bias control trials (catch trials) were present in the location experiment and involved a fake binocular-rivalry stimulus purposely manipulated with a spatial mix of half red and half green. The inclusion of such catch trials is standard practice when using binocular rivalry to measure the sensory trace of mental imagery (e.g., Keogh & Pearson, 2017; Pearson et al., 2008). These trials monitored whether participants were veridically reporting binocular-rivalry dominance and assessed the presence of any possible nonperceptual bias effects (e.g., task demands or intentional participant misreporting). All participants correctly reported a mixed response during these trials (M > 90% correct), demonstrating that it is unlikely that the results were driven by nonperceptual bias, including demand characteristics and intentional participant misreporting.

Discussion

Here, we used a perceptual illusion, binocular rivalry, to measure the sensory trace of emergent thoughts during attempts to control and suppress them. Our data suggest that when individuals attempt to suppress or substitute a thought, a sensory representation, or trace, of the thought forms. This sensory trace appears to emerge even when individuals report successful thought suppression, suggesting the intriguing possibility that these sensory representations, possibly nonconscious, may trigger thought-control failure. In addition, we found thought control to be worse in highly anxious and schizotypic individuals but more successful in mindful individuals.

Binocular-rivalry priming for suppressed thoughts was significantly above chance and, surprisingly, was comparable with priming for actively imagined thoughts (see Fig. 2a). This suggests that the sensory representations of suppressed visual thoughts were as strong as those of imagined thoughts and that thought suppression was thus ineffective in eliminating the sensory trace of these thoughts. These results reflect the paradoxical nature of thought suppression (Wegner et al., 1987): Active suppression ironically led to the formation of a sensory representation of the unwanted visual thought. The bias from suppressed thoughts showed the sensory characteristics of low-level visual areas, as indicated by their susceptibility to changes in luminance and spatial location of the rivalry stimulus. This suggests that suppressed visual thoughts are represented in visual areas, including early visual cortices (Boynton et al., 1996; Goodyear & Menon, 1998). Strikingly, even when participants reported successful thought suppression, subsequent rivalry was still significantly biased, suggesting that the perceptual effects of emergent sensory representations were adequate to bias rivalry but not trigger a subjective report of suppression failure. This is important because trials in which suppression was believed to be successful composed nearly 70% of the total number of suppression trials (see Fig. S1), indicating, for the majority of trials, that individuals believed they were successful in thought suppression.

It is important to recognize that the observed results were sensory in nature and were not due to high-level decisional biases, such as demand characteristics, participant misreporting, or semantic priming. First, all participants correctly responded on decision-bias and memory control trials, indicating that they were veridically reporting rivalry dominance and properly attending to the task. Second, our results showed that perceptual manipulations (uniform luminance and retinotopic location) modulated both priming and the probability of thought-suppression failure (see Fig. 2). Because these low-level perturbations occur only at the level of early visual cortices, this provides strong evidence that the priming effects and self-reported thought-suppression failure we observed were driven from the sensory visual cortex and not by any demand, expectation, or intentional misreporting. Third, because participants were explicitly informed that there was no right or wrong response and, thus, they should respond truthfully, we believe it is unlikely that participants were intentionally misreporting their responses, according to task demands or otherwise (e.g., misreporting of thought-control success or failure), or conducting the experiment dishonestly (e.g., secretly using a thought-substitution strategy instead of a thought-suppression strategy in the thought-suppression experiment). Verbal reports by participants after the experiment corroborated this. Lastly, the binocular-rivalry paradigm used in this study has been established as a valid and reliable method in multiple published studies over the past decade to objectively measure visual mental imagery (e.g., Chang et al., 2013; Keogh & Pearson, 2011, 2017, 2018; Pearson et al., 2008; Sherwood & Pearson, 2010). This includes ruling out possible nonsensory explanations such as task demands or participant misreporting in explaining observed results.

Thought substitution, in which participants substituted the thought of a red or green object with a white object, was more effective in controlling visual thoughts compared with thought suppression. Priming for substituted thoughts was significantly lower than both imagined and suppressed thoughts and was not significantly different from chance (see Fig. 2a). This suggests that thought substitution may work to neutralize the emergent sensory trace of the unwanted thought. However, although these results reflect the effectiveness of employing thought substitution to improve thought control (Wegner et al., 1987), thought substitution was not completely effective. When thought-substitution attempts failed, the priming effect returned (see Fig. 2b), indicating that the strategy was effective only during successful substitution. However, individuals showed good metacognition of when their thought-substitution attempts were successful or failed. Although more effective than thought suppression, thought substitution failed, on average, 13% of the time (see Fig. S1), showing that this strategy did not completely eliminate the probability of thought-control failure.

Our results reveal that individuals vary in their ability to control visual thoughts. We found that the average thought-control-index score was not significantly above zero (see Fig. S5), indicating that individuals were, on average, poor at controlling their visual thoughts using thought suppression. Specifically, individuals with high anxiety and schizotypy showed the lowest levels of thought control, whereas mindful individuals exhibited the highest levels of thought control (see Fig. 3).

Interestingly, individuals with high trait anxiety or schizotypy showed negative thought-control-index scores, implying that the sensory strength of suppressed thoughts was even stronger than that of intentionally imagined thoughts. This suggests that deficiencies in thought control for anxious or schizotypic individuals may arise, in part, from a lack of control over the sensory representations of visual thoughts. Indeed, the inability to control mental imagery is known to play a significant role in anxiety disorders and schizophrenia (Hirsch & Holmes, 2007; Oertel et al., 2009). On the other hand, individuals with high levels of mindfulness showed positive thought-control-index scores, suggesting that the sensory strength of suppressed thoughts was weaker than that of imagined thoughts. This reflects research showing the effectiveness of mindfulness in improving mental control (Brown & Ryan, 2003).

Our study offers a novel and unique method to objectively measure thought control using binocular rivalry. By probing the sensory basis of emergent thoughts, we have demonstrated a way to examine the underlying mechanisms of thought control and its failure. In addition, our measure enables examination into the metacognition of thought control as well as into thoughts that exist in the absence of subjective awareness. This paradigm could be important in objectively measuring thought control, a domain that has thus far been difficult to investigate because of its inherent subjectivity.

Future research should explore the underlying neural mechanism behind thought control and its failure, including its metacognition, and why such mechanisms go awry in particular populations. Such mechanistic understanding will provide new avenues for clinical treatments. In sum, this article introduces the first objective sensory-based method to investigate thought control. We provide data to suggest that thought-control failure may emerge from involuntary, and possibly nonconscious, sensory representations that arise during attempts at thought control and that show significant individual differences across psychological traits.

Supplemental Material

KwokSupplementalMaterial_rev – Supplemental material for Measuring Thought-Control Failure: Sensory Mechanisms and Individual Differences

Supplemental material, KwokSupplementalMaterial_rev for Measuring Thought-Control Failure: Sensory Mechanisms and Individual Differences by Eugene L. Kwok, Gaelle Leys, Roger Koenig-Robert and Joel Pearson in Psychological Science

Supplemental Material

Kwok_OpenPracticesDisclosure_rev – Supplemental material for Measuring Thought-Control Failure: Sensory Mechanisms and Individual Differences

Supplemental material, Kwok_OpenPracticesDisclosure_rev for Measuring Thought-Control Failure: Sensory Mechanisms and Individual Differences by Eugene L. Kwok, Gaelle Leys, Roger Koenig-Robert and Joel Pearson in Psychological Science

Footnotes

Acknowledgements

We would like to thank Rebecca Keogh and Sebastian Rogers for their valuable help and advice.

Action Editor

Philippe G. Schyns served as action editor for this article.

Author Contributions

E. L. Kwok and G. Leys are joint first authors of this article. All the authors developed the study concept and methods. G. Leys and R. Koenig-Robert wrote the code for the experiment and analysis. G. Leys conducted the pilot experiments. G. Leys and E. L. Kwok carried out testing and data collection. E. L. Kwok and R. Koenig-Robert conducted the statistical analysis. E. L. Kwok wrote the manuscript. R. Koenig-Robert and J. Pearson provided critical revisions and comments. All the authors approved the final manuscript for submission.

ORCID iD

E. L. Kwok

Declaration of Conflicting Interests

The author(s) declared that there were no conflicts of interest with respect to the authorship or the publication of this article.

Funding

This work was supported by the Australian National Health and Medical Research Council (Grant Nos. APP1024800, APP1046198, and APP1085404), a Career Development Fellowship (Grant No. APP1049596 to J. Pearson), and an Australian Research Council Discovery Project (Grant No. DP140101560).

Supplemental Material

Additional supporting information can be found at

Open Practices

All data have been made publicly available via the Open Science Framework and can be accessed at osf.io/b69g8. The design and analysis plans for this study were not preregistered. The complete Open Practices Disclosure for this article can be found at https://journals-sagepub-com.web.bisu.edu.cn/doi/suppl/10.1177/0956797619837204. This article has received the badge for Open Data. More information about the Open Practices badges can be found at .

References

Alais

Blake

(2005). Binocular rivalry. Cambridge, MA: MIT Press.

Benjamini

Hochberg

(1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society B: Methodological, 57, 289–300.

Boynton

G. M.

Engel

S. A.

Glover

G. H.

Heeger

D. J.

(1996). Linear systems analysis of functional magnetic resonance imaging in human V1. The Journal of Neuroscience, 16, 4207–4221.

Brainard

D. H.

(1997). The Psychophysics Toolbox. Spatial Vision, 10, 433–436.

Brown

K. W.

Ryan

R. M.

(2003). The benefits of being present: Mindfulness and its role in psychological well-being. Journal of Personality and Social Psychology, 84, 822–848.

Chang

Lewis

D. E.

Pearson

(2013). The functional effects of color perception and color imagery. Journal of Vision, 13(10), Article 4. doi:10.1167/13.10.4

Chapman

L. J.

Chapman

J. P.

(1973). Disordered thought in schizophrenia. East Norwalk, CT: Appleton-Century-Crofts.

Chiou

Rich

A. N.

Rogers

Pearson

(2018). Exploring the functional nature of synaesthetic colour: Dissociations from colour perception and imagery. Cognition, 177, 107–121.

Foa

E. B.

Huppert

J. D.

Leiberg

Langner

Kichic

Hajcak

Salkovskis

P. M.

(2002). The Obsessive-Compulsive Inventory: Development and validation of a short version. Psychological Assessment, 14, 485–496.

10.

Freeston

M. H.

Dugas

M. J.

Ladouceur

(1996). Thoughts, images, worry, and anxiety. Cognitive Therapy and Research, 20, 265–273.

11.

Giuliano

R. J.

Wicha

N. Y.

(2010). Why the white bear is still there: Electrophysiological evidence for ironic semantic activation during thought suppression. Brain Research, 1316, 62–74.

12.

Goodyear

B. G.

Menon

R. S.

(1998). Effect of luminance contrast on BOLD fMRI response in human primary visual areas. Journal of Neurophysiology, 79, 2204–2207.

13.

Hirsch

C. R.

Holmes

E. A.

(2007). Mental imagery in anxiety disorders. Psychiatry, 6, 161–165.

14.

Hubel

D. H.

Wiesel

T. N.

(1959). Receptive fields of single neurones in the cat’s striate cortex. The Journal of Physiology, 148, 574–591.

15.

Keogh

Pearson

(2011). Mental imagery and visual working memory. PLOS ONE, 6(12), Article e29221. doi:10.1371/journal.pone.0029221

16.

Keogh

Pearson

(2014). The sensory strength of voluntary visual imagery predicts visual working memory capacity. Journal of Vision, 14(12), Article 7. doi:10.1167/14.12.7

17.

Keogh

Pearson

(2017). The perceptual and phenomenal capacity of mental imagery. Cognition, 162, 124–132.

18.

Keogh

Pearson

(2018). The blind mind: No sensory visual imagery in aphantasia. Cortex, 105, 53–60.

19.

Kosslyn

S. M.

Brunn

Cave

K. R.

Wallach

R. W.

(1984). Individual differences in mental imagery ability: A computational analysis. Cognition, 18, 195–243.

20.

Kulik

C. T.

Perry

E. L.

Bourhis

A. C.

(2000). Ironic evaluation processes: Effects of thought suppression on evaluations of older job applicants. Journal of Organizational Behavior, 21, 689–711.

21.

Lavy

E. H.

van den Hout

M. A.

(2009). Thought suppression induces intrusions. Behavioural Psychotherapy, 18, 251–258.

22.

Lin

Y. J.

Wicker

F. W.

(2007). A comparison of the effects of thought suppression, distraction and concentration. Behaviour Research and Therapy, 45, 2924–2937.

23.

Mason

Claridge

Jackson

(1995). New scales for the assessment of schizotypy. Personality and Individual Differences, 18(1), 7–13.

24.

Mitchell

J. P.

Heatherton

T. F.

Kelley

W. M.

Wyland

C. L.

Wegner

D. M.

Neil Macrae

(2007). Separating sustained from transient aspects of cognitive control during thought suppression. Psychological Science, 18, 292–297.

25.

Oertel

Rotarska-Jagiela

van de Ven

Haenschel

Grube

Stangier

. . . Linden

D. E.

(2009). Mental imagery vividness as a trait marker across the schizophrenia spectrum. Psychiatry Research, 167, 1–11.

26.

Pearson

(2014). New directions in mental-imagery research: The binocular-rivalry technique and decoding fMRI patterns. Current Directions in Psychological Science, 23, 178–183.

27.

Pearson

Clifford

C. W.

Tong

(2008). The functional impact of mental imagery on conscious perception. Current Biology, 18, 982–986.

28.

Pearson

Naselaris

Holmes

E. A.

Kosslyn

S. M.

(2015). Mental imagery: Functional mechanisms and clinical applications. Trends in Cognitive Sciences, 19, 590–602.

29.

Petrie

K. J.

Booth

R. J.

Pennebaker

J. W.

(1998). The immunological effects of thought suppression. Journal of Personality and Social Psychology, 75, 1264–1272.

30.

Purdon

Clark

D. A.

(2002). The need to control thoughts. In Frost

R. O.

Steketee

(Eds.), Cognitive approaches to obsessions and compulsions: Theory, assessment, and treatment (pp. 29–43). Amsterdam, The Netherlands: Elsevier.

31.

Rachman

de Silva

(1978). Abnormal and normal obsessions. Behaviour Research and Therapy, 16, 233–248.

32.

Sherwood

Pearson

(2010). Closing the mind’s eye: Incoming luminance signals disrupt visual imagery. PLOS ONE, 5(12), Article e15217. doi:10.1371/journal.pone.0015217

33.

Slotnick

S. D.

Thompson

W. L.

Kosslyn

S. M.

(2005). Visual mental imagery induces retinotopically organized activation of early visual areas. Cerebral Cortex, 15, 1570–1583.

34.

Spielberger

C. D.

(1983). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.

35.

Stone

A. A.

(2000). The science of self-report: Implications for research and practice. Mahwah, NJ: Erlbaum.

36.

Wegner

D. M.

(1989). White bears and other unwanted thoughts: Suppression, obsession, and the psychology of mental control. London, England: Penguin Press.

37.

Wegner

D. M.

(1994). Ironic processes of mental control. Psychological Review, 101, 34–52.

38.

Wegner

D. M.

Schneider

D. J.

Carter

S. R.

III White

T. L.

(1987). Paradoxical effects of thought suppression. Journal of Personality and Social Psychology, 53, 5–13.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.84 MB

0.58 MB