Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation

Abstract

Pain-related cognitive biases have been demonstrated in chronic pain patients, yet despite theoretical predictions are rarely investigated in combination. Combined cognitive biases were explored in individuals with chronic headache (n = 17) and pain-free controls (n = 20). Participants completed spatial cueing (attentional bias), sentence generation (interpretation bias) and free recall tasks (memory bias), with ambiguous sensory-pain, disability and neutral words. Individuals with chronic headache, relative to controls, showed significantly greater interpretation and memory biases favouring ambiguous sensory-pain words and interpretation bias favouring ambiguous disability words. No attentional bias was found. Further research is needed exploring the temporal pattern of cognitive biases.

Keywords

chronic illness cognitive processing health psychology pain quantitative methods

Introduction

Cognitive biases are frequently explored in people with chronic pain, with evidence for pain-related attentional (Crombez et al., 2013; Schoth et al., 2012), interpretation (Schoth and Liossi, 2016) and memory biases (Pincus and Morley, 2001). Rather than existing in isolation, however, different forms of bias may influence and interact with one another (Hirsch et al., 2006). Schema theory (Beck et al., 1985; Beck and Haigh, 2014) predicts dysfunctional schemata result in a selective processing of schema-congruent information, with biases existing across various forms of information processing that are largely considered in parallel. Within the depression literature, it has been speculated that attentional allocation and rumination over negative information precede a negative interpretation bias, with evidence found for an indirect effect between attentional and memory biases, with interpretation bias as a mediating variable (Everaert et al., 2013). Considering pain, the Threat Interpretation Model (Todd et al., 2015) highlights the ambiguity of certain words used in attentional bias paradigms (e.g. sharp-pain, sharp-cleaver) and argues that an interpretation bias favouring pain-related meanings is necessary, but not sufficient, for attentional bias to be observed. Although the co-existence of cognitive biases is therefore often mentioned in the literature, theoretical perspectives differ in their conceptualisation of the sequential nature of biases.

Research with chronic pain patients typically explores cognitive biases in isolation, with no study to date assessing attentional, interpretation and memory biases in the same sample. Of the few studies exploring two forms of biases, one found evidence for both interpretation and memory biases for illness-related words, although the proportion of illness-related interpretations made was not correlated with recall scores (Pincus et al., 1996). Another study reported memory bias for sensory-pain words, but not attentional bias measured via the emotional Stroop task; correlations between attentional and memory biases were not reported (Pincus et al., 1998). Comparison between these studies is difficult due to the different pairs of cognitive bias explored and the different stimuli categories used. Related to this latter issue, it is uncertain whether combined cognitive biases are shown for pain-related information specifically or also for more general disability-related information. Research using the visual probe task has found pronounced attentional biases in individuals with chronic pain towards sensory-pain words but not disability words (e.g. Dehghani et al., 2003; Sharpe et al., 2009), whereas one study reported interpretation biases for both sensory-pain and disability words (McKellar et al., 2003). While some studies have found memory biases for sensory-pain words (e.g. Serbic and Pincus, 2014), none to date have used a category of disability words specifically.

The aim of this study was to provide a preliminary investigation of attentional, interpretation and memory biases for disorder-specific sensory-pain words and more general disability words in individuals with chronic headache. Considering attentional biases in chronic pain are particularly pronounced at longer stimuli presentation times associated with rumination (Crombez et al., 2013; Schoth et al., 2012), attentional biases were assessed prior to interpretation biases. Based on the theoretical models highlighted and the results of former research, it was hypothesised that individuals with chronic headache, relative to pain-free controls, would show significant attentional, interpretation and memory biases for sensory-pain words.

Methods

Ethics approval was obtained from the University of Southampton Research Ethics Committee. In accordance with regulations of the University and guidelines of the Declaration of Helsinki, participants provided informed consent prior to taking part.

Participants

Participants were recruited from the South of England via press advertisements. For the chronic headache group, inclusion criteria were as follows: (a) suffering from primary tension-type headache or migraine and satisfying the criteria stated in the International Classification of Headache Disorders, 3rd edition, beta version (ICHD-3); (b) aged 18 years or over; and (c) having normal or corrected-to-normal vision. Exclusion criteria were as follows: (a) having a diagnosis or receiving treatment for any psychiatric disorder, either currently or within the past 5 years, and (b) suffering from any other form of chronic pain including secondary headache. For the control group, inclusion criteria were as follows: (a) aged 18 years or over and (b) having normal or corrected-to-normal vision. Exclusion criteria were as follows: (a) having a diagnosis or receiving treatment for any psychiatric disorder, either currently or within the past 5 years; (b) suffering from any form of chronic or regular pain (in terms of headache frequency, more than seven headaches per month); and (c) taking any psychotropic or analgesic medication regularly. Eligibility was established via telephone interview prior to recruitment.

A total of 37 participants were recruited, including 17 with chronic headache (mean age = 38.76, standard deviation (SD) = 13.66, range 21–59) and 20 healthy controls (mean age = 35.55, SD = 13.78, range: 21–62). The majority of participants were female (30; 81%). Participants with chronic headache reported on average 21.53 (SD = 6.37, range: 15–30 days) headache days per month, while healthy controls reported on average 1.95 (SD = 1.61, range: 0–5 days) headache days per month. Participants with chronic headache reported living with chronic headache for a mean duration of 14.12 (SD = 11.40, range: 1–35) years. Within this group, 6 (35.3%) had tension-type headache and 11 (64.7%) had migraine. Nine (52.9%) reported at least one relative to also suffer from regular headache. As indexed by their Migraine Disability Assessment (MIDAS) scores, 11 (64.7%) indicated severe disability as a consequence of their headaches. All but three (82.4%) were taking regular analgesic medication for the management of their headache.

Measures

The following questionnaires were used: Hospital Anxiety and Depression Scale (Zigmond and Snaith, 1983), State–Trait Anxiety Inventory (Spielberger et al., 1970), McGill Pain Questionnaire (Melzack, 1975), Brief Pain Inventory–Short Form (Cleeland and Ryan, 1994) and MIDAS Questionnaire (Stewart et al., 2001). Full details are provided in Supplementary Questionnaire Information and Data.

Experimental stimuli

Experimental stimuli (Supplementary Table 1 available at: http://hpq.sagepub.com/) included nine sensory-pain words, reflecting the sensory dimension of headache pain; nine disability words, reflecting long-lasting potential consequences of pain; and nine neutral words, unrelated specifically to pain or ill health. As interpretation biases were explored, all words were either homographs (i.e. words which have identical spelling but different meanings and etymologies) or pseudo-homographs (also referred to as polysemes, that is, words which have identical spelling but different meanings, although stem historically from the same source) (Drury, 1969). All words therefore have multiple meanings and associations, as indexed by the Merriam-Webster Dictionary (merriam-webster.com) and the University of Florida Free Association Norms database (Nelson et al., 2004). Sensory-pain and disability words have neutral associations as well as pain-related/disability associations, and neutral words have multiple neutral associations. Words across the three conditions were matched on length and Kucera–Francis written frequency using the MRC Psycholinguistic Database (Wilson, 1988). Using the University of Florida Free Association Norms database, words were also matched on word set size, which is an index of how many strong associations the word has (Nelson et al., 2004).

Experimental paradigms

Spatial cueing task

The exogenous spatial cueing task (Supplementary Figure 1 available at: http://hpq.sagepub.com/) was modelled on those used in recent studies exploring biases in depressed and pain populations (Baert et al., 2010; Everaert et al., 2013; Martin and Chapman, 2010). The task included 216 experimental trials, presented in a new randomised order for each participant. The task began with eight practice trials featuring nonsense strings of consonants (e.g. Ghtxyw). Experimental trials were split into two 108 trial blocks, each of which was immediately preceded by two buffer trials (featuring nonsense strings). Each trial began with a fixation cross in the centre of the screen for 1000 ms, flanked to the left and right by two rectangular boxes. A single word (size 40 Times New Roman font) was then presented for 100 or 1500 ms in either the left or right box. Fifty milliseconds after the disappearance of the word, a probe (the cue) (1 cm in diameter) appeared in the left or right box. Participants indicated the location of the probe as quickly and accurately as possible using a two-button response box to provide their responses. Trials followed one another automatically, with the inter-trial interval varying randomly between 1000 and 1500 ms. If a response was not provided after 2000 ms of the probe being displayed, the trial ended automatically.

Trials in which the probe replaces the word are ‘valid’ trials, and trials in which the probe appears in the opposite location to the word are ‘invalid’ trials. An equal number of valid and invalid trials were included, divided equally across the three word conditions and two presentation times. Words were therefore not predictive of the probe location. Participants were instructed to focus on the fixation cross as much as possible; in reality, however, the presentation of the word is difficult to ignore and typically engages attention (Fox et al., 2001), the extent of which can be compared across participant groups and stimuli categories. Six catch trials (three per block) were also included and interspersed randomly among the experimental trials, which reinforced the instruction to focus on the centre of the screen. During each catch trial, a number (1, 2 or 3) replaced the fixation cross. Participants pressed the corresponding number key on the keyboard as soon as they saw the number and then placed their thumbs again on the buttons on the response box. Following each catch trial, the subsequent experimental trial began automatically after 2500 ms. Text, fixation crosses and probes were presented in white against a black background. The spatial cueing task lasted approximately 20 minutes.

Sentence generation task

The sentence generation task (Taghavi et al., 2000) included 27 experimental trials featuring the same sensory-pain, disability and neutral stimuli as the spatial cueing task (Supplementary Figure 1 available at: http://hpq.sagepub.com/). Trials were presented in a new randomised order for each participant. Each trial began with a fixation cross for 1000 ms. A single word presented in size 40 Times New Roman font then replaced the cross, which remained on the screen until the end of the trial. Participants were instructed to read the word and using the keyboard type a single sentence featuring the word once only. Text appeared in size 18 Times New Roman font below the experimental word as the participant typed. Backspace and delete keys were used to correct spelling mistakes or make amendments as necessary, and when satisfied the participant pressed the F12 key to submit their response. If no answer was submitted, the next trial began after 120 seconds. Trials followed one another automatically, and all 27 trials were presented in a single block. Two practice trials were initially presented to familiarise participants with the requirements of the task, featuring the words running and dancing. Text, fixation crosses and cursors were presented in black against a white background. The sentence generation task lasted approximately 12 minutes.

Free recall task

Participants were unexpectedly given 3 minutes to write down as many words as possible from the spatial cueing and sentence generation tasks.

Apparatus and procedure

Spatial cueing and sentence generation tasks were developed in Presentation^® (version 12.2; Neurobehavioural Sciences) and run on a personal computer with a 15-in monitor. Participants first completed the spatial cueing task, and after a short break completed the sentence generation task. This task order is the same as that used in a recent study exploring combined cognitive biases in subclinical depression (Everaert et al., 2013) and reflects the notion that attentional allocation and rumination over negative/threatening information precede a bias favouring negative/threatening interpretations of ambiguous information. Immediately following this, participants were instructed to count backwards from 400 in units of 7 for 2 minutes. This distractor task was used to reduce the possibility of recency effects influencing subsequent recall. Participants then completed the surprise free recall task. After a short break, participants completed the questionnaires, which were presented in a new randomised order for each participant. The total experimental duration was approximately 60 minutes.

Data reduction and analytic plan

Analyses were conducted in IBM SPSS Statistics for Windows 22. For the spatial cueing task, practice trials were excluded from final analysis, along with experimental trials with incorrect responses. Box and whisker plots for overall data revealed outliers to be any response latencies falling below 200 or above 1500 ms, which were removed. After this, mean response times were calculated for each participant individually, with any response >3SD away from their individual mean also removed as outliers. Response times from valid trials were subtracted from invalid trials to form a cue validity index for each stimuli category at both presentation times (Koster et al., 2010). An overall attentional bias index was then calculated by subtracting the cue validity index of neutral words from the cue validity index of sensory-pain and disability words at 100 and 1500 ms presentation times (Everaert et al., 2013). A positive score indicates greater attentional engagement of sensory-pain/disability words relative to neutral words, whereas a negative value indicates greater attentional engagement of neutral words relative to sensory-pain/disability words. This attentional bias index was used in the statistical analyses.

For the sentence generation task, as per former research (McKellar et al., 2003), two raters independently and blindly categorised participant response sentences as pain-related (i.e. describing the experience of pain – He had a pressing pain in his head), disability (i.e. describing the consequences of pain or illness – His lack of mobility was a pressing matter) or benign (i.e. describing situations or events unrelated to pain or disability – The boy was pressing the buttons in the lift). Benign responses include both neutral and positive sentences. The initial inter-rater agreement was 97 per cent, and after discussion, consensus was reached on 100 per cent of ratings. The proportion of interpretations made was used in the analysis (Pincus et al., 1994, 1996). For the free recall task, the proportion of words recalled per stimuli category was computed (Karimi et al., 2016; Pincus et al., 1996).

Between-group differences for demographic characteristics and self-report questionnaires were explored via t-tests and χ² for continuous and categorical variables, respectively. Mixed-designs analysis of variance (ANOVA) was used to compare chronic headache and healthy control groups on attentional bias indices, sentence generation responses and proportion of words recalled. ANOVAs and t-tests were used as required in post hoc analyses to clarify significant effects. Effect sizes for ANOVA and t-tests were quantified using partial eta-square ηp² and Cohen’s d, respectively. Cohen’s d and associated 95 per cent confidence intervals (CIs) were calculated via exploratory software for confidence intervals (ESCI) (Cumming, 2012). For ANOVA analyses, the alpha level was set at .05, two-tailed. Pearson’s correlation coefficients were conducted selectively to assess the relationship between different types of sensory-pain and disability cognitive bias.

Results

Group comparisons

Chronic headache (n = 17) and healthy control groups (n = 20) did not differ significantly in age (t(35) = 0.71, p = .482, d = 0.23, CI of d (−0.42, 0.88)) or sex (chronic headache = 94% female (16); healthy control = 70% female (14), χ² = 3.48, p = .062). Questionnaire data are provided in Supplementary Questionnaire Information and Data. Independent t-tests were conducted on measures completed by both groups. The chronic headache group reported significantly higher trait anxiety than the healthy control group. Trait anxiety was not included as a covariate in the analyses conducted as this was not predictive of cognitive bias for sensory-pain or disability words.

Spatial cueing task: attentional bias

Mean reaction times across the three stimuli conditions did not differ significantly, F(2, 72) = 2.08, p = .133, ηp² = .055 (sensory-pain = 624 ms, disability = 619 ms, and neutral = 621 ms). Chronic headache and healthy control groups did not differ significantly in mean reaction time (chronic headache = 664.97 ms (SD = 139.00); healthy control = 584.63 ms (SD = 120.18); t(35) = 1.89, p = .068, d = 0.62, CI of d (−0.05, 1.28)); number of incorrect responses made (chronic headache = 2.24 (SD = 1.72); healthy control = 2.10 (SD = 2.17); t(35) = 0.21, p = .837, d = 0.07,CI of d (−0.58, 0.72)); or number of outliers removed (chronic headache = 4.06 (SD = 3.21); healthy control = 3.10 (SD = 1.62); t(35) = −1.17, p = .249, d = 0.39, CI of d (−0.27, 1.04)).

Attentional bias index scores are presented in Table 1. A 2 (group: chronic headache, healthy control) × 2 (stimuli type: sensory-pain, disability) × 2 (presentation time: 100 ms, 1500 ms) mixed-designs ANOVA was conducted on attentional bias scores. No significant main effects or interactions were found, for example: group, F(1, 35) = 0.22, p = .639, ηp² = .006; group by stimuli type by presentation time, F(1, 35) = 2.62, p = .115, ηp² = .070. Although the lack of significant results did not warrant the comparison of groups via independent-samples t-tests, Cohen’s d effect sizes for between-group differences are given here for information purposes: sensory-pain 100 ms d = 0.13, CI of d (−0.52, 0.77); sensory-pain 1500 ms d = 0.19, CI of d (−0.46, 0.83); disability 100 ms d = 0.28, CI of d (−0.37, 0.93); and disability 1500 ms d = 0.16, CI of d (−0.49, 0.80). Full results are presented in Supplementary Table 2 available at: http://hpq.sagepub.com/.

Table 1.

Attentional, interpretation and memory bias mean scores (SD) for chronic headache and healthy control groups.

Bias index	Chronic headache group (n = 17)	Healthy control group (n = 20)
Attentional bias (bias index scores)
Sensory-pain, 100 ms	16.14 (64.38)	9.79 (33.60)
Sensory-pain, 1500 ms	4.97 (42.95)	−1.88 (30.73)
Disability, 100 ms	17.41 (74.09)	1.79 (34.79)
Disability, 1500 ms	−5.17 (47.19)	1.70 (40.20)
Interpretation bias (proportion of responses)
Pain responses to sensory-pain words	.275 (.153)	.114 (.116)
Disability responses to sensory-pain words	.000 (.000)	.000 (.000)
Benign responses to sensory-pain words	.726 (.153)	.886 (.116)
Pain responses to disability words	.022 (.048)	.022 (.046)
Disability responses to disability words	.419 (.180)	.310 (.127)
Benign responses to disability words	.560 (.183)	.668 (.124)
Pain responses to neutral words	.000 (.000)	.006 (.025)
Disability responses to neutral words	.020 (.044)	.028 (.049)
Benign responses to neutral words	.980 (.044)	.967 (.063)
Memory bias (proportion of words recalled)
Sensory-pain words	.469 (0.177)	.324 (0.138)
Disability words	.318 (0.149)	.370 (0.155)
Neutral words	.213 (0.144)	.306 (0.153)

SD: standard deviation.

Sentence generation task: interpretation bias

Chronic headache and healthy control groups did not differ significantly in the number of valid interpretations made (chronic headache = 26.65 (SD = 0.70); healthy control = 26.55 (SD = 1.00); t(35) = −0.34, p = .739, d = 0.11, CI of d (−0.54, 0.76)). The proportion of pain, disability and benign interpretations made for each category of words is presented in Table 1. A 2 (group: chronic headache, healthy control) × 3 (stimuli type: sensory-pain, disability, and neutral) × 3 (response: pain, disability, and benign) mixed-designs ANOVA was conducted on the proportion of participant responses classified as pain, disability or benign for each stimulus category. The main effect of response was significant, F(2, 55) = 1208.79, p < .001, ηp² = .972, along with a significant group by response interaction, F(2, 35) = 10.18, p < .001, ηp² = .225, and stimuli type by response interaction, F(4, 140) = 102.04, p < .001, ηp² = .745. These results were qualified by a significant group by stimuli type by response interaction, F(4, 140) = 6.34, p = .002, ηp² = .153.

Independent t-tests were conducted to clarify the significant three-way interaction. Chronic headache participants, relative to healthy controls, provided significantly more pain responses to sensory-pain words, t(35) = 3.64, p = .001, d = 1.19, CI of d (0.48, 1.88), and significantly more disability responses to disability words, t(35) = 2.14, p = .040, d = 0.71, CI of d (0.03, 1.37). In contrast, healthy controls, relative to chronic headache participants, provided significantly more benign responses to sensory-pain words, t(35) = 3.63, p = .001, d = 1.20, CI of d (0.49, 1.90), and significantly more benign responses to disability words, t(35) = 2.11, p = .042, d = 0.71, CI of d (0.04, 1.37). No significant differences between groups were found for the neutral stimuli category.

Free recall task: memory bias

Chronic headache and healthy control groups did not differ significantly in the number of total words recalled (chronic headache = 8.88 (SD = 2.62); healthy control = 9.25 (SD = 3.31); t(35) = 0.370, p = .714, d = 0.12, CI of d (−0.53, 0.77)) or the number of incorrect words recalled (chronic headache = 0.65 (SD = 1.06); healthy control = 0.70 (SD = 1.22); t(35) = 0.14, p = .890, d = 0.05, CI of d (−0.60, 0.69)). The proportion of the total correct recall accounted for by each stimuli condition is provided in Table 1. A 2 (group: chronic headache, healthy control) × 3 (stimuli type: sensory-pain, disability, and neutral) mixed-designs ANOVA was conducted. The main effect of stimuli type was significant, F(2, 70) = 5.00, p = .009, ηp² = .125, as was the group by stimuli-type interaction, F(2, 70) = 4.27, p = .018, ηp² = .109. Independent t-tests were conducted to clarify the significant interaction. Chronic headache participants, relative to healthy controls, recalled a significantly greater proportion of sensory-pain words, t(35) = 2.81, p = .008, d = 0.92, CI of d (0.24, 1.60). There was a trend for healthy controls, relative to chronic headache participants, to recall more neutral words, although this was not statistically significant, t(35) = 1.90, p = .066, d = 0.63, CI of d (−0.04, 1.28). The two groups did not differ in the proportion of disability words recalled, t(35) = 1.04, p = .306, d = 0.35, CI of d (−0.31, 1.00). One-way ANOVAs were also conducted on stimuli type for each group independently. A main effect of stimuli type was found for the chronic headache group only, F(2, 32) = 7.61, p = .002, ηp² = .322. Pairwise comparisons revealed a significantly greater proportion of words recalled were from the sensory-pain category than the neutral category (mean difference = .256, p = .006).

Correlation analysis

Correlation matrices can be found in Supplementary Tables 3(a) to (f) available at: http://hpq.sagepub.com/. Across all participants, the proportion of pain interpretations made for sensory-pain words was positively correlated with the proportion of sensory-pain words recalled, r = .391, p = .017. For the healthy control group, the proportion of pain interpretations made for sensory-pain words was positively correlated with attentional biases for sensory-pain words presented at 100 ms, r = .500, p = .025.

Discussion

Partly supporting the study hypothesis, individuals with chronic headache, relative to controls, showed significantly greater interpretation and memory biases for sensory-pain words. The two groups did not differ in patterns of attentional bias for sensory-pain words. The results also revealed a significant interpretation bias for disability words in those with chronic headache relative to controls.

Interpretation biases have been infrequently explored in chronic pain, although all former studies showed significantly more frequent pain-related/illness-related interpretations of ambiguous words or images in those with chronic pain relative to pain-free controls (Schoth and Liossi, 2016). The specificity of such biases is a notable issue, as some studies used broader stimuli categories reflecting ill health rather than pain specifically (e.g. Pincus et al., 1994, 1996). The present results provide some clarification, showing that individuals with chronic headache demonstrate interpretation biases for both disorder-specific sensory-pain words and general disability words. The majority of participants experienced pain for many years and reported severe disability, and therefore, biases for disability words are unsurprising. These results also align with those of McKellar and colleagues, who reported pain- and disability-related interpretation biases in individuals with chronic pain. In contrast, studies using the visual probe task have reported attentional biases for sensory-pain but not disability words (e.g. Dehghani et al., 2003; Sharpe et al., 2009). This is supported by a meta-analysis showing significant biases for sensory-pain words, but not words associated with consequences of pain (including disability-related words), in chronic pain patients (Crombez et al., 2013). Furthermore, research recruiting a chronic headache sample with similar levels of disability found attentional bias for pain-related images but not general health-threat images (Schoth and Liossi, 2013). This pattern of results therefore suggests that interpretation biases exist for a broader range of stimuli than do attentional biases, although it is premature at present to try and explain why this should be so.

Evidence for sensory-pain memory biases in chronic pain patients have been reported in some studies (e.g. Serbic and Pincus, 2014) but not others (e.g. Busch et al., 2006). Although disability-related words have not been used specifically, broader categories of words reflecting ill health have been used, again with some studies reporting memory biases (e.g. Serbic and Pincus, 2014) and others not (e.g. Nikendei et al., 2009). Numerous methodological differences exist within this literature which likely account for the inconsistency of results, including whether the recall task is made explicit or is unexpected by participants. If unexpected, there is also variation in the conditions and instructions under which encoding takes place. Interestingly, this study found evidence of a pronounced memory bias for sensory-pain, but not disability words, in those with chronic headache. Within-group analysis also showed that a significantly greater proportion of words recalled by the chronic headache group were from the sensory-pain category than the neutral category, although no difference was found between disability and neutral categories. While a specificity of memory bias was observed, future research should aim to replicate these findings under a variety of experimental conditions, including explicit and unexpected tasks.

Attentional bias was measured via the spatial cueing task. As this paradigm presents one stimulus per trial, it may be addressed whether the presence of threat has a slowing or inhibition effect on motor responses (Mogg et al., 2008); overall, there was no evidence of such effects in this study. Stimuli presentation times of 100 and 1500 ms were used to explore initial orienting of attention and maintained attention, respectively. While the former is associated with hypervigilance for threat (Beck et al., 1985), the latter has been linked to processes of excessive elaboration and rumination (Donaldson et al., 2007). No evidence of attentional bias for sensory-pain words was found, however, which is inconsistent with former studies recruiting chronic headache samples that have reported pain-related biases via visual probe (e.g. Liossi et al., 2009; Schoth and Liossi, 2013), visual scanning (Liossi et al., 2014) and visual search tasks (Schoth et al., 2015). The spatial cueing task has been infrequently used in pain samples, with evidence for pain-related biases in one study with patients with irritable bowel syndrome (Chapman and Martin, 2011) but not another (Martin and Chapman, 2010). A recent study reported evidence of attentional biases towards health and somatic threat cues in low-pain catastrophisers, but not high-pain catastrophisers as predicted (Schrooten et al., 2015). Furthermore, a review of research with anxious populations showed a very small, non-significant between-group effect size for studies using the spatial cueing task (Bar-Haim et al., 2007), and between-group effect sizes were also small in this study. Related to this, for between-group comparisons, a post hoc power calculation using G*Power (Faul et al., 2007) revealed only 9, 31 and 65 per cent probabilities of correctly rejecting the null hypothesis for small (0.2) medium (0.5) and large (0.8) effect sizes, respectively. The paradigm used, in combination with the small sample size, likely explains the null attentional bias findings in this study.

As two experimental conditions were included in this study, a neutral match was required for each experimental word. The spatial cueing task was therefore used instead of the visual probe task, as the latter would necessitate either twice as many neutral words or the same set of neutral words paired with both sensory-pain and disability words. Both are problematic for the later exploration of memory biases, which requires an equal number of words per category, shown the same number of times, to allow for a valid comparison of results (i.e. if neutral words are displayed twice as often, memory for this category of words will likely improve). This problem is avoided by the use of the spatial cueing task, which presents words from each stimuli category the same number of times. However, considering the null findings, follow-up studies will explore combined cognitive biases using the visual probe task with a single experimental stimuli category of sensory-pain words (matched with neutral words).

The Threat Interpretation Model (Todd et al., 2015) predicts a relationship between pain-related interpretation and attentional biases. For healthy controls, an interpretation bias favouring the pain-related meaning of ambiguous words with sensory-pain and neutral meanings is associated with a hypervigilance for such words. Surprisingly, this effect was not found for the chronic headache group. Across all participants, the proportion of pain interpretations made for sensory-pain words was positively correlated with the proportion of sensory-pain words recalled. As pain is important for survival (Williams, 2002), it is unsurprising that across all individuals a tendency to interpret ambiguous information as pain-related is associated with enhanced recall of such information. However, specific testing of the predictions raised by the Threat Interpretation Model is required, although this latter correlation does highlight the possibility of modifying this model to also include memory biases.

Further research is required to explore the temporal pattern of cognitive biases, especially as the effects of attentional and interpretation biases on the experience of pain are being explored (Carleton et al., 2011; Jones and Sharpe, 2014; Schoth et al., 2013; Sharpe et al., 2012). While Todd and colleagues predict interpretation biases to precede attentional biases, schema theory largely considers cognitive biases in parallel (Beck and Haigh, 2014), and the combined cognitive bias hypothesis states that ‘a number of biased cognitive processes often operate simultaneously and/or in succession’ (Hirsch et al., 2006: 223). As noted, research exploring combined biases in subclinical depression postulates that attentional allocation and rumination over negative information precede a negative interpretation bias (Everaert et al., 2013). Counterbalancing the order of attentional and interpretation bias tasks in future research would help elucidate the precise relationship between these forms of cognitive bias. Longitudinal research may also be used to address whether induction of one form of bias has resulting effects on other forms of bias (Hirsch et al., 2006; Schoth and Liossi, 2016).

Limitations may be highlighted with the current investigation. Due to the small sample size, the study was underpowered, and the results of the spatial cueing task and the correlational analysis should be interpreted with caution. The study also did not explore the relationship between pain characteristics, such as intensity and duration of episodes, on patterns of cognitive bias. While there is inconsistent evidence for such relationships in the existing literature, there is a need to explore the impact of such characteristics in future combined cognitive bias research. Furthermore, and as noted, this study did not set out to specifically test the predictions raised by the Threat Interpretation Model (Todd et al., 2015). Rather, the aim was to provide a preliminary investigation into combined cognitive biases in chronic pain, an area which has received little attention. In conclusion, the results contribute to a growing body of chronic pain research, providing evidence for interpretation and memory pain-related biases in individuals with chronic headache.

Supplemental Material

Supplementary_Figure_1 – Supplemental material for Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation

Supplemental material, Supplementary_Figure_1 for Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation by Daniel E Schoth, Laura Parry and Christina Liossi in Journal of Health Psychology

Supplemental Material

Supplementary_Questionnaire_Information_and_Data – Supplemental material for Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation

Supplemental material, Supplementary_Questionnaire_Information_and_Data for Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation by Daniel E Schoth, Laura Parry and Christina Liossi in Journal of Health Psychology

Supplemental Material

Supplementary_Tables_3a_-_3f._Correlations – Supplemental material for Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation

Supplemental material, Supplementary_Tables_3a_-_3f._Correlations for Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation by Daniel E Schoth, Laura Parry and Christina Liossi in Journal of Health Psychology

Supplemental Material

Supplementary_Table_1._Stimuli – Supplemental material for Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation

Supplemental material, Supplementary_Table_1._Stimuli for Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation by Daniel E Schoth, Laura Parry and Christina Liossi in Journal of Health Psychology

Supplemental Material

Supplementary_Table_2._Attentional_bias_ANOVA_results – Supplemental material for Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation

Supplemental material, Supplementary_Table_2._Attentional_bias_ANOVA_results for Combined cognitive biases for pain and disability information in individuals with chronic headache: A preliminary investigation by Daniel E Schoth, Laura Parry and Christina Liossi in Journal of Health Psychology

Footnotes

Acknowledgements

The authors would like to thank Dr Jin Zhang, PhD (Senior Experimental Officer, Academic Unit of Psychology, University of Southampton, Southampton, Hampshire, UK) for developing the spatial cueing and sentence generation tasks used in this investigation and for her continued technical support.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

References

Baert

De Raedt

Schacht

. (2010) Attentional bias training in depression: Therapeutic effects depend on depression severity. Journal of Behavior Therapy and Experimental Psychiatry 41: 265–274.

Bar-Haim

Lamy

Lee

. (2007) Threat-related attentional bias in anxious and nonanxious individuals: A meta-analytic study. Psychological Bulletin 133: 1–24.

Beck

Haigh

(2014) Advances in cognitive theory and therapy: The generic cognitive model. Annual Review of Clinical Psychology 10: 1–24.

Beck

Emery

Greenberg

(1985) Anxiety Disorders and Phobias: A Cognitive Perspective. New York: Basic Books.

Busch

Montgomery

Melin

. (2006) Visuospatial and verbal memory in chronic pain patients: An explorative study. Pain Practice 6: 179–185.

Carleton

Richter

Asmundson

GJG

(2011) Attention modification in persons with fibromyalgia: A double blind, randomized clinical trial. Cognitive Behaviour Therapy 40: 279–290.

Chapman

Martin

(2011) Attention to pain words in irritable bowel syndrome: Increased orienting and speeded engagement. British Journal of Health Psychology 16: 47–60.

Cleeland

Ryan

(1994) Pain assessment: Global use of the brief pain inventory. Annals of the Academy of Medicine, Singapore 23: 129–138.

Crombez

Van Ryckeghem

Eccleston

. (2013) Attentional bias to pain-related information: A meta-analysis. Pain 154: 497–510.

10.

Cumming

(2012) Understanding the New Statistics: Effect Sizes, Confidence Intervals, and Meta-analysis. New York: Routledge.

11.

Dehghani

Sharpe

Nicholas

(2003) Selective attention to pain-related information in chronic musculoskeletal pain patients. Pain 105: 37–46.

12.

Donaldson

Lam

Mathews

(2007) Attention and rumination in major depression. Behaviour Research and Therapy 45: 2664–2678.

13.

Drury

(1969) Homographs and pseudo-homographs. Word Ways: The Journal of Recreational Linguistics 2: 146–154.

14.

Everaert

Tierens

Uzieblo

. (2013) The indirect effect of attention bias on memory via interpretation bias: Evidence for the combined cognitive bias hypothesis in subclinical depression. Cognition & Emotion 27: 1450–1459.

15.

Faul

Erdfelder

Lang

A-G

. (2007) G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods 39: 175–191.

16.

Fox

Russo

Bowles

. (2001) Do threatening stimuli draw or hold visual attention in subclinical anxiety? Journal of Experimental Psychology: General 130: 681–700.

17.

Hirsch

Clark

Mathews

(2006) Imagery and interpretations in social phobia: Support for the combined cognitive biases hypothesis. Behavior Therapy 37: 223–236.

18.

Jones

Sharpe

(2014) The effect of Cognitive Bias Modification for Interpretation (CBM-I) on avoidance of pain during an acute experimental pain task. Pain 155: 1569–1576.

19.

Karimi

Pilenko

Held

. (2016) Recall bias in patients with chronic low back pain: Individual pain response patterns are more important than pain itself! International Journal of Behavioral Medicine 23: 12–20.

20.

Koster

De Raedt

Leyman

. (2010) Mood-congruent attention and memory bias in dysphoria: Exploring the coherence among information-processing biases. Behaviour Research and Therapy 48: 219–225.

21.

Liossi

Schoth

Bradley

. (2009) Time course of attentional bias for pain-related cues in chronic daily headache sufferers. European Journal of Pain 13: 963–969.

22.

Liossi

Schoth

Godwin

. (2014) Using eye movements to investigate selective attention in chronic daily headache. Pain 155: 503–510.

23.

McKellar

Clark

Shriner

(2003) The cognitive specificity of associative responses in patients with chronic pain. British Journal of Clinical Psychology 42: 27–39.

24.

Martin

Chapman

SCE

(2010) Cognitive processing in putative functional gastrointestinal disorder: Rumination yields orientation to social threat not pain. European Journal of Pain 14: 207–213.

25.

Melzack

(1975) The McGill Pain Questionnaire: Major properties and scoring methods. Pain 1: 277–299.

26.

Mogg

Holmes

Garner

. (2008) Effects of threat cues on attentional shifting, disengagement and response slowing in anxious individuals. Behaviour Research and Therapy 46: 656–667.

27.

Nelson

McEvoy

Schreiber

(2004) The University of South Florida free association, rhyme, and word fragment norms. Behavior Research Methods, Instruments, & Computers 36: 402–407.

28.

Nikendei

Waldherr

Schiltenwolf

. (2009) Memory performance related to organic and psychosocial illness attributions in somatoform pain disorder patients. Journal of Psychosomatic Research 67: 199–206.

29.

Pincus

Morley

(2001) Cognitive-processing bias in chronic pain: A review and integration. Psychological Bulletin 127: 599–617.

30.

Pincus

Fraser

Pearce

(1998) Do chronic pain patients ‘Stroop’ on pain stimuli? British Journal of Clinical Psychology 37: 49–58.

31.

Pincus

Pearce

Perrott

(1996) Pain patients’ bias in the interpretation of ambiguous homophones. British Journal of Medical Psychology 69: 259–266.

32.

Pincus

Pearce

McClelland

. (1994) Interpretation bias in responses to ambiguous cues in pain patients. Journal of Psychosomatic Research 38: 347–353.

33.

Schoth

Liossi

(2013) Specificity and time-course of attentional bias in chronic headache: A visual-probe investigation. Clinical Journal of Pain 29: 583–590.

34.

Schoth

Liossi

(2016) Biased interpretation of ambiguous information in patients with chronic pain: A systematic review and meta-analysis of current studies. Health Psychology 35(9): 944–956.

35.

Schoth

Delgado Nunes

Liossi

(2012) Attentional bias towards pain-related information in chronic pain; a meta-analysis of visual-probe investigations. Clinical Psychology Review 32: 13–25.

36.

Schoth

Georgallis

Liossi

(2013) Attentional bias modification in people with chronic pain: A proof of concept study. Cognitive Behaviour Therapy 42: 233–243.

37.

Schoth

Godwin

Liversedge

. (2015) Eye movements during visual search for emotional faces in individuals with chronic headache. European Journal of Pain 19: 722–732.

38.

Schrooten

Vancleef

Vlaeyen

(2015) Attention allocation to ambiguous health/somatic threat cues. European Journal of Pain 19: 1002–1011.

39.

Serbic

Pincus

(2014) Diagnostic uncertainty and recall bias in chronic low back pain. Pain 155: 1540–1546.

40.

Sharpe

Dear

Schrieber

(2009) Attentional biases in chronic pain associated with rheumatoid arthritis: Hypervigilance or difficulties disengaging? The Journal of Pain 10: 329–335.

41.

Sharpe

Ianiello

Dear

. (2012) Is there a potential role for attention bias modification in pain patients? Results of 2 randomised, controlled trials. Pain 153: 722–731.

42.

Spielberger

Gorsuch

Lushene

(1970) State Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.

43.

Stewart

Lipton

Dowson

. (2001) Development and testing of the Migraine Disability Assessment (MIDAS) Questionnaire to assess headache-related disability. Neurology 56: S20–S28.

44.

Taghavi

Moradi

Neshat-Doost

. (2000) Interpretation of ambiguous emotional information in clinically anxious children and adolescents. Cognition & Emotion 14:809–822.

45.

Todd

Sharpe

Johnson

. (2015) Towards a new model of attentional biases in the development, maintenance, and management of pain. Pain 156: 1589–1600.

46.

Williams

(2002) Facial expression of pain: An evolutionary account. Behavioral and Brain Sciences 25: 439–455.

47.

Wilson

(1988) MRC psycholinguistic database: Machine-usable dictionary, version 2.00. Behavior Research Methods, Instruments, & Computers 20: 6–10.

48.

Zigmond

Snaith

(1983) The hospital anxiety and depression scale. Acta Psychiatrica Scandinavica 67: 361–370.

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

0.03 MB

0.01 MB

0.02 MB