If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
High and low fear-avoidant chronic low back pain (CLBP) patients were examined.
•
CLBP patients imagined back-straining activities during functional magnetic resonance imaging.
•
High versus low fear-avoidant CLBP patients showed stronger hippocampus activations.
•
High fear-avoidant CLPB patients versus pain-free control subjects did not differ.
Abstract
The fear-avoidance model postulates that in an initial acute phase chronic low back pain (CLBP) patients acquire a fear of movement that results in avoidance of physical activity and contributes to the pain becoming chronic. The current functional magnetic resonance imaging study investigated the neural correlates of imagining back-straining and neutral movements in CLBP patients with high (HFA) and low fear avoidance (LFA) and healthy pain-free participants. Ninety-three persons (62 CLBP patients, 31 healthy controls; age 49.7 ± 9.2 years) participated. The CLBP patients were divided into an HFA and an LFA group using the Tampa Scale of Kinesiophobia. The participants viewed pictures of back-straining and neutral movements and were instructed to imagine that they themselves were executing the activity shown. When imagining back-straining movements, HFA patients as well as healthy controls showed stronger anterior hippocampus activity than LFA patients. The neural activations of HFA patients did not differ from those of healthy controls. This may indicate that imagining back-straining movements triggered pain-related evaluations in healthy controls and HFA participants, but not in LFA participants. Although heightened pain expectancy in HFA compared with LFA patients fits well with the fear-avoidance model, the difference between healthy controls and LFA patients was unexpected and contrary to the fear-avoidance model. Possibly, negative evaluations of the back-straining movements are common but the LFA patients use some kind of strategy enabling them to react differently to the back-straining events.
Perspective
It appears that low fear-avoidant back pain patients use some kind of strategy or underlying mechanism that enables them to react with less fear in the face of potentially painful movements. This warrants further investigation because countering fear and avoidance provide an important advantage with respect to disability.
and may culminate in a vicious cycle characterized by pain → catastrophizing → fear of movement/hypervigilance → avoidance of movement → pain sensitization → pain.
This behavioral pattern may lead to increasing disability and, if the avoidance is extended to various areas of life, can give rise to generalized withdrawal and abet the development of a depressive syndrome.
Fear of movement and avoidance behaviour toward physical activity in chronic-fatigue syndrome and fibromyalgia: State of the art and implications for clinical practice.
In the past decade, the FA model has generated a large amount of research activity, accruing evidence that supports the model with regard to chronic low back pain (CLBP)
However, it has also been recognized that not all CLBP patients show a fear of movement, but that there may be a subgroup of patients who react with endurance rather than avoidance when faced with pain.
Pain-related avoidance versus endurance in primary care patients with subacute back pain: psychological characteristics and outcome at a 6-month follow-up.
Despite the FA model's pervasiveness, little is known about the neural correlates of the postulated fear. In an earlier study we investigated the neural correlates of viewing back-straining movements with an event-related functional magnetic resonance imaging (fMRI) study
and expected to find neural activities in structures related to fear processing, such as the amygdala, insula, cingulate gyrus, fusiform gyrus, and the substantia nigra.
However, we failed to find any neural correlates of fearful processing of such pictures of back-straining movements in high fear-avoidant CLBP participants, despite the fact that their fear processing in general was unimpaired.
One major limitation of our previous study was that the participants only passively viewed the photographs of the movements and thus it could not be excluded that they may have focused on other aspects of the pictures and failed to relate the movement to themselves. To address this limitation, the present study specifically instructed the participants to imagine that they themselves were performing the movements shown in the pictures. Moreover, we tested a much larger sample (n = 93) and included men.
We used an fMRI block design to compare CLBP patients with HFA and LFA and healthy pain-free participants when they imagined executing back-straining and neutral movements. We hypothesised that when patients with high fear-avoidance (HFA) imagined back-straining movements as opposed to neutral movements, they would show activations in regions related to fear processing. We also hypothesised that these activations would be more pronounced for patients with HFA than for patients with LFA or for controls.
Methods
Participants
Recruitment
Patients with CLBP were recruited via local rehabilitation centers and newspaper advertisements. Patients were included if the location of their back pain was the lower back, the pain had persisted for at least 6 months and they reported an average pain intensity of at least 5 on a numerical rating scale from 0 to 10 (with 0 being ‘no pain’ and 10 being the ‘worst imaginable pain’) over the previous 4 weeks. Specific causes of their back pain had been ruled out by their physicians. Healthy controls were recruited via newspaper advertisements. All of the participants were screened for further pain conditions, neurological disorders, and standard magnetic resonance imaging (MRI) contraindications, such as metal objects in the body. In addition, the German version of the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition Axis I Disorders (SCID I)
was conducted with each participant to exclude any mental disorders.
Sample Characteristics
Ninety-three persons (mean age 49.7 ± 9.2 years) participated in the study. Of these participants, 62 suffered from unspecific CLBP and 31 were pain-free controls matched for age and sex (HC). The participants with CLBP were divided by means of a median split performed on the basis of their scores in the German version of the Tampa Scale of Kinesiophobia (TSK)
Pain ratings on a numerical rating scale ranging from 0 (‘no pain’) to 10 (‘worst pain imaginable’) over the previous 4 weeks.
7.6
1.0
7.9
1.7
1.01
-
PDI mean
2.7
1.7
3.6
2.1
1.75
-
PCS
15.5
11.5
18.4
9.9
1.03
-
NOTE. LFA indicates chronic low back pain patients with low fear-avoidance; HFA, chronic low back pain patients with high fear-avoidance; and HC, healthy pain-free control participants. A χ2 test showed no difference for sex between the groups (χ2 = .26, df = 2, P > .05). For CES-D, Tukey post hoc tests showed: HC was lower than both CLBP groups, all P < .01. For STAI-trait, Tukey post hoc tests showed: HFA was higher than HC, P < .05.
*P < .05.
**P < .01.
***P < .001.
† Pain ratings on a numerical rating scale ranging from 0 (‘no pain’) to 10 (‘worst pain imaginable’) over the previous 4 weeks.
A 3 × 2 repeated measures design with the between-factor group (HFA, LFA, and HC) and the within-factor movement type (back-straining and neutral) was used.
Psychometric Instruments
Pain Characteristics and Fear of Movement
Pain duration, location, and intensity as well as other pain characteristics were assessed using the pain questionnaire of the German chapter of the International Association for the Study of Pain (Deutscher Schmerzfragebogen).
Nagel B, Gerbershagen HU, Lindena G, Pfingsten M: Development and empirical validation of the German pain questionnaire of the DGSS [in German]. Pain 16:263-270, 2002
The participants rated the average, maximum, and minimum pain intensity they had experienced in the previous 4 weeks on a numerical rating scale from 0 to 10 (with 0 being ‘no pain’ and 10 representing the ‘worst imaginable pain’). Pain-related disability was assessed using the German version of the Pain Disability Index (PDI).
The PDI measures subjective disability (ie, the extent to which the chronic pain interferes with the person's ability to engage in everyday activities). The instrument consists of 7 items to be rated from 0 to 10 (with 0 being ‘no disability’ and 10 being ‘total disability’), which cover the following areas: family/home responsibilities; recreation; social activities; occupation; sexual behavior; self-care and life-support activities. Catastrophizing cognitions were assessed using the German version of the Pain Catastrophizing Scale (PCS).
Pain catastrophizing denotes the tendency to experience pain as very threatening and to exaggerate its seriousness. The PCS consists of 13 statements regarding what the person may think or feel when they are in pain (example item: ‘It's terrible and I think it's never going to get any better’) and the participants have to judge their agreement with each statement on a 5-point rating scale (0 = ‘not at all’ to 4 = ‘all the time’).
The TSK is a reliable instrument for the assessment of fear of movement in CLBP patients, with satisfactory internal consistency (Cronbach α = .76–.84)
and consists of 17 statements expressing fear of physical activity (eg, ‘I'm afraid that I might injure myself if I exercise’), which have to be rated on a 4-point rating scale from 1 = ‘strongly disagree’ to 4 = ‘strongly agree.’ The scales for items 4, 8, 12, and 16 are inverted.
Depression and Anxiety
Manifest depression was excluded through the SCID I interview. To measure depression as a dimensional construct, we used the Center for Epidemiologic Studies Depression Scale (CES-D).
Hautzinger M: The CES-D Scale: An instrument to measure depression in investigations of the general population [in German]. Diagnostica 34:167–173, 1988
The CES-D is well suited to assessing depression in clinical and nonclinical populations and consists of 20 items pertaining to various possible symptoms of depression (example item: ‘I felt depressed’). For each item, the participants had to rate on a 4-point scale the number of days during the previous week that he or she had experienced the symptom being described (0 = ‘rarely’ [less than 1 day], 1 = ‘sometimes’ [1–2 days], 2 = ‘often’ [3–4 days], 3 = ‘most of the time’ [5–7 days]). The items 4, 8, 12, and 16 are scored inversely.
Trait and state anxiety were measured with the German version of the State-Trait-Anxiety Inventory (STAI).
The state version consists of 20 items describing thoughts or feelings related to fearful, worried states (example item: ‘I feel tense’), for which the person has to rate on a scale from 1 = ‘not at all’ to 4 = ‘very much so’ how much the statement describes how they presently feel. Half of the items (1, 2, 5, 8, 10, 11, 15, 16, 19, 20) are scored inversely. The trait version also consists of 20 items describing a fearful and worried disposition (example item: ‘I worry too much over something that really does not matter’). The person has to rate on a scale from 1 = ‘almost never’ to 4 ‘almost always’ whether these statements describe how they feel in general. The items (1, 6, 7, 10, 13, 16, 19) are scored inversely.
Stimulus Material
A total of 48 color photographs of movements and postures served as stimulus material. The pictures were identical to the movement pictures we used in our previous study.
supplemented by pictures produced by one of the authors (A.B.). The photographs were selected from a larger pool of 138 pictures on the basis of a pilot study with 21 CLBP patients and 20 controls. For more detail of the picture selection process see Barke et al.
Twenty-four photographs each depicted either back-straining or neutral movements and postures (Fig 1). The back-straining movements showed activities that CLBP patients normally avoid, such as bending down or lifting heavy boxes (numbers of the PHODA pictures used: 1, 2, 3, 4, 20, 22, 29, 31, 46, 57, 64, 65, 68, 76, 82, 83, 85, 87, 88, 91, 99). The neutral movements showed less strenuous activities, such as sitting, standing, or walking (they also included a few PHODA pictures: 8, 60, 61, 69, 84, 92).
Figure 1Example pictures of back-straining movement (left) and a neutral movement (right). The left picture is part of the PHODA,
The study protocol was approved by the Ethics Committee of the University Medical Center, Göttingen. Before their participation, the participants received full written information about the study and signed an informed consent form. The participants received a small amount of financial compensation for their participation in the study and a CD with the anatomical pictures of their brains.
All participants attended 2 separate appointments. During the first appointment, the SCID I interview was carried out and the participants received the questionnaires, which they filled in at home and brought with them for the second appointment. During the second appointment, the MRI scanning took place. Immediately before the MRI session, the participants filled in the state version of the STAI to assess their anxiety before the scan. The participants were then placed in the MRI scanner in a supine position, wearing foam earplugs for noise protection and headphones for communication with the experimenter and additional noise protection. After the anatomical reference scans, the functional imaging commenced and the pictures were presented to the participants using magnetic resonance compatible LCD goggles with a resolution of 800 × 600 (Resonance Technology, Northridge, CA). If a participant required glasses, corrective lenses were combined with the goggles to ensure corrected-to-normal vision.
Stimulation Paradigm
The pictures were presented in alternating (back-straining/neutral) blocks. Each block lasted 24 seconds and contained 4 pictures, presented for 6 seconds each, either of back-straining movements (back-straining) or neutral movements (neutral). Twelve back-straining and 12 neutral blocks were shown, resulting in a total duration of 9.6 minutes. Each picture was presented twice. The picture order was randomised except for the block structure. The participants were instructed to look at the pictures and to imagine that they themselves were executing the activity shown in the picture.
Image Acquisition
The MRI scans were acquired at 3 Tesla (Siemens Magnetom TIM Trio; Siemens Healthcare, Erlangen, Germany) with an 8-channel phased-array head coil. For anatomical reference, a 3-D T1-weighted data set was acquired (turbo fast low angle shot; echo time = 3.26 ms; repetition time = 2,250 ms; inversion time = 900 ms; flip angle = 12°; isotropic resolution = 1 × 1 × 1 mm3). The functional data sets were acquired using T2*-weighted gradient-echo echo-planar imaging (echo time = 36 ms, repetition time = 2000 ms, flip angle = 70°, 22 slices of 4 mm thickness at an in-plane resolution of 2 × 2 mm2). A total of 288 whole-brain volumes were recorded within the functional run. To reach magnetic saturation, 4 preparatory scans were acquired and subsequently discarded from the analyses.
Stimulus Ratings
After the scanning session, the participants were asked to view the pictures again, this time on a computer screen, and to rate each movement in terms of valence and arousal on a 9-point scale using the computerized version of the self-assessment manikin.
For valence the instruction was: ‘Please rate: how positive/negative is the depicted movement for you personally?’; for arousal it was: ‘Please rate: how strong is the arousal you feel with respect to the depicted movement?’ The end points of the valence scale were marked as ‘very negative’ and ‘very positive,’ and the end points of the arousal scale were ‘none’ and ‘very strong.’ Each picture was presented separately and the order of presentation was randomized. The picture remained visible until the participant had rated it and pressed the button labelled ‘next.’ The ratings of 2 participants were not recorded because of a technical problem. The imagining was not measured directly, because doing so would have disturbed the act of imagining the actions. However, at the end of the experiment we interviewed the participants as to whether they found it hard to imagine the movements in the timeframe and whether they had succeeded in doing so. The answers indicated that the participants had indeed imagined the depicted actions and postures (and nearly all participants recalled spontaneously some of the movements telling us how they went about it).
Data Analysis
Functional Data
The functional data were analyzed using Brain Voyager QX Software version 2.1.2 (Brain Innovation, Maastricht, The Netherlands). Standard preprocessing steps included 3-D motion correction, slice scan time correction, temporal filtering (linear trend removal and high-pass filtering) and spatial smoothing with a Gaussian kernel (full width at half maximum 8 × 8 × 8 mm3). For high pass filtering and linear trend removal, the presence of low frequencies was estimated and removed using a general linear model containing sines and cosines (up to 2 cycles per run, .0034 Hz). The filtering was performed in the time domain using the general linear model to estimate and remove the contributions of low frequencies in each voxel's time course. The functional data sets were coregistered to the anatomical reference scans and transformed into Talairach space. A brain mask was generated on the basis of the Talairach-transformed anatomic images. Subsequent functional analyses were restricted to the voxels in this mask. Group analysis was performed using the multisubject (random effects) approach of the general linear model. In a first step, the 24-second duration of each block (boxcar) was convolved with the canonical hemodynamic response function resulting in 2 predictors (back-straining movements, neutral movements) for the types of movement. On the basis of these predictors, β values were estimated for each voxel, which express the strength of the predictor for that voxel. This step included a percentage normalization of the raw blood oxygen level-dependent signal time-courses to account for differences in mean blood oxygen level-dependent signal levels across voxels or subjects. In a first step, β maps for the contrast (back-straining > neutral) were generated for each participant. In the second-level analysis, the β maps were used to examine the effect of movement type across all participants, regardless of group membership, and for each group individually. For these analyses, the Bonferroni-corrected results are reported. In a second step, the groups were compared in planned contrasts and the contrasts between HFA > LFA, HFA > HC, and LFA > HC are reported. For the group comparisons, the uncorrected cluster threshold was set at P = .001. On the basis of the number of activated voxels and the estimated smoothness of the map, Monte-Carlo simulations (1,000 iterations) were performed to determine the minimum cluster size required to yield an error rate of no more than P < .05 at the cluster level. Activations were assigned to anatomical locations according to the nearest coordinates in the Talairach Daemon database.
Often, clusters span several structures. For these, we report the cluster maxima. In addition, further activated structures are included in the tables whenever a local maximum of the t-score was located in the structure. A local maximum is defined as a maximum at least 10 mm apart from the next maximum. If 2 local maxima are closer together, only the maximum with the higher statistical value is reported and the other is suppressed. For each structure, only the maximum with the highest statistical value is reported.
For a correlation analysis, CLBP patients with HFA and LFA were pooled and voxelwise correlations with the TSK score were calculated. The uncorrected threshold was set at P = .001 and cluster level corrections applied.
All clusters quoted in this study are numbered consecutively.
Questionnaire Data and Stimulus Ratings
Questionnaire data pertaining to the participants of all 3 groups (CES-D and STAI) as well as age were compared between the groups using 1-way analysis of variance (ANOVA), using Tukey post hoc tests, when appropriate. For measures applicable to the participants with back pain only (TSK, PDI, PCS, pain duration, pain intensity) independent t-tests (2-tailed) were used to compare the high and low fear-avoidant participants. As measures of effect size, Cohen's d and η2 were calculated.
With regard to the valence and arousal ratings collected after the scanning session, we calculated two 3 × 2 repeated measures ANOVAs for valence and arousal of the movements with the between-subject factor group (HFA, LFA, HC) and the within-subject factor movement type (back-straining, neutral). In cases in which Mauchly test indicated that the sphericity assumption had been violated, Greenhouse–Geisser corrections were performed and the resulting values reported. Tukey post hoc tests were used for further analyses.
Results
Stimulus Ratings
Valence and Arousal Ratings for the Neutral and Back-Straining Movements
Valence
The 3 × 2 repeated measures ANOVA for the valence ratings of the movements with the between-subjects factor group and the within-subjects factor movement type showed no main effect for group (P > .10), but revealed a main effect for movement type (F1,88 = 368.06, P < .001, η2 = .629) and an interaction group × movement type (F2,88 = 5.16, P < .005, η2 = .018). Tukey post hoc tests showed that, within each group, the back-straining movements were rated more negatively than the neutral movements (all Ps < .001). Tukey tests found no significant differences between the neutral movements between the groups. For the back-straining movements, the HFA and the LFA CLBP patients rated the pictures as more negative than the HC (all Ps < .01), but did not differ from each other (P > .10; see Fig 2 for more detail).
Figure 2Valence and arousal ratings for the stimuli as a function of group and movement type. Error bars show standard deviations. For the valence ratings, low numbers indicate negative valence (1 = very negative to 9 = very positive). For the arousal ratings, low numbers indicate low levels of arousal (1 = no arousal to 9 = extreme arousal). HFA indicates high fear-avoidant chronic low back pain patients; LFA, low fear-avoidant chronic low back pain patients; and HC, healthy pain-free participants.
The 3 × 2 repeated measures ANOVA for the arousal ratings showed main effects for group (F2,88 = 5.95, P < .005, η2 = .070) and movement type (F1,88 = 106.61, P < .001, η2 = .213), and an interaction Group × Movement (F2,88 = 5.29, P < .01, η2 = .021). Tukey post hoc tests showed that, within each group, the back-straining movements were rated as more arousing than the neutral movements (all Ps < .005). Tukey post hoc tests for group showed that HFA pain patients rated the movements as more arousing than HC (P < .01) and so did LFA patients (P < .01); however, no difference was observed between HFA and LFA. Post hoc tests for group × movement showed that the HFA as well as the LFA rated the back-straining movements as more arousing than the HC (HFA: P < .001; LFA: P < .05), but did not differ from each other. The neutral pictures were rated equally nonarousing by all groups (see Fig 2 for more detail).
Imaging Results
Back-Straining Movements > Neutral Movements
All participants
Across all groups imagining back-straining movements evoked stronger activations than imagining neutral movements in the right middle temporal gyrus (1; cluster numbers in parentheses), the right inferior parietal lobule (3), the bilateral superior parietal lobules (2, 4), the left middle frontal gyrus (6), and the left precentral gyrus (5). These structures were the locations of the clusters’ peak voxels. However, the clusters extended to adjacent structures (see Table 2 for a complete list). The inverse contrast (neutral movements > back-straining movements) activated the left posterior cingulate cortex (8) and the right cuneus (7).
Table 2Structures Activated for the Contrast Back-Straining Movements > Neutral Movements and the Reverse Contrast Across All Groups (N = 93; Bonferroni Corrected; Total Area Activated: 104,557 mm3)
When HFA participants imagined executing back-straining movements as opposed to neutral movements, the right middle occipital gyrus (9, 10), the left inferior temporal gyrus (15), the bilateral superior parietal lobuli (11, 14), the left cuneus (13), and areas in the right cerebellum (12) were activated (see Table 3 for a complete list and Fig 3 middle column). The inverse contrast evoked activations in the right cuneus (16).
Table 3Structures Activated for the Contrast Back-Straining Movements > Neutral Movements and the Reverse Contrast Within the HFA (Bonferroni-Corrected; Total Area Activated 13,884 mm3)
Figure 3Bonferroni-corrected activation maps for the within-group contrast back-straining movements > neutral movements. Left column: healthy pain-free participants; middle column: high fear-avoidant chronic low back pain patients; right column: low fear-avoidant chronic low back pain patients. The statistical maps are shown overlaid on the averaged T1-weighted data set of all subjects.
In the LFA group, imagining executing back-straining movements activated the right fusiform gyrus (17) and the left inferior temporal gyrus (18). The inverse contrast evoked activations in the bilateral posterior cingulate (19, 21) and the right cuneus (20) (see Table 4 for a complete list and Fig 3 right column).
Table 4Structures Activated for the Contrast Back-Straining Movements > Neutral Movements and the Reverse Contrast Within the LFA (Bonferroni Corrected; Total Area Activated 6,739 mm3)
When HC participants imagined carrying out back-straining movements as opposed to neutral movements, the bilateral occipital gyri (22, 25), the right inferior and superior parietal lobule (23, 24), the left precuneus (27) and cuneus (26), the left inferior temporal gyrus (29), and the declive (cerebellum) (28) were activated, whereas the inverse contrast evoked activations in the right cuneus (30) (see Table 5 for a complete list and Fig 3, left column).
Table 5Structures Activated for the Contrast Back-Straining Movements > Neutral Movements and the Reverse Contrast Within the HC (Bonferroni Corrected; Total Area Activated 25,269 mm3)
Comparing HFA patients imagining back-straining movements as opposed to neutral movements with LFA patients, we found that the HFA showed stronger activations in the right hippocampus (32), the right parahippoccampal gyrus (31), and the left middle temporal gyrus (33). The inverse contrast (LFA > HFA) was empty (Table 6 and Fig 4, left column). For the clusters identified by the contrast HFA > LFA we calculated the β differences back-straining movements > neutral movements for all groups, including the control group (Fig 5).
Table 6Structures Activated for the Group Comparisons HFA > LFA and HC > LFA (P < .001, Cluster Level Corrected; All Other Group Comparisons Were Empty)
Figure 4Second-level comparisons (between group) of the contrast back-straining movements > neutral movements. The statistical maps were cluster-level corrected and are shown overlaid on the averaged T1-weighted data set of all subjects. HFA indicates high fear-avoidant chronic low back pain patients; LFA, low fear-avoidant chronic low back pain patients; and HC, healthy pain-free participants.
Figure 5β Values of the predictors for the clusters 31–33 (A–C). Please note that the clusters were extracted on the basis of the contrast HFA > LFA (shown in black) and the values for HC (shown in grey) are only provided for purposes of comparison. HFA indicates high fear-avoidant chronic low back pain patients; LFA, low fear-avoidant chronic low back pain patients; and HC, healthy pain-free participants.
No differences were found between the back pain participants with HFA and the pain-free control participants.
LFA > HC
For the contrast LFA > HC, no activations were found. Inversely, pain-free controls showed stronger activations in the right hippocampus (34) and the right lingual gyrus (35) than LFA patients (see Table 6 and Fig 4, right column).
Correlations With the TSK
Combining the back pain patients into 1 group (HFA and LFA: n = 62) and calculating the correlation of the contrast back-straining movements > neutral movements with the TSK score showed no regions above the threshold of P < .001.
Discussion
We investigated the neural correlates of imagining back-straining movements in high and low fear-avoidant CLBP patients using fMRI and compared them with those of pain-free controls. Patients with LFA differed from patients with HFA and pain-free controls, but contrary to expectations, patients with HFA and control participants did not show any differences in their neural correlates. On a behavioral level, back pain patients rated the back-straining movements as more negative and more arousing than did pain-free control participants.
The main effect for movement type, imagining back-straining movements relative to neutral movements, evoked activations in areas that Caspers and colleagues identified in their meta-analysis
as consistently involved in action observation and imitation (eg, middle occipital gyrus, fusiform gyrus, superior parietal lobule, precentral gyrus). Imagined actions are represented similarly to overtly performed and observed actions.
Because back-straining movements (eg, lifting a heavy-looking box) were more complex activities than neutral movements (eg, standing, sitting), this seems to indicate that the participants indeed followed the instructions and observed and imagined the presented movements.
When they imagined back-straining movements in contrast to neutral movements, patients with HFA showed stronger activations in the right anterior (ventral) hippocampus, the right parahippocampal gyrus, and the left middle temporal gyrus than the low fear-avoidant patients. The hippocampus receives afferents from the septum and the area entorhinalis; its efferents reach the thalamus and the gyrus cingulum and it is reciprocally connected with the basal amygdala.
Evidence of anatomical and functional variability along the hippocampal long axis has inspired different proposals of long-axis functional specialization.
In particular, the anterior hippocampus was shown to play an important role in anxiety: in animal studies, lesions in the ventral hippocampus were consistently associated with reduced anxiety in a number of standard anxiety paradigms.
Activation in the right hippocampus was positively correlated (r = .74) with high pain expectancy sensitivity during the expectancy of a painful thermal stimulation,
and the hippocampus found to be differentially activated in response to identical noxious stimuli, depending on whether the perceived pain was enhanced by pain-related anxiety. In the context of the present study, anterior hippocampus activation may indicate that imagining back-straining movements in contrast to neutral movements triggered pain-related representations to arise in high fear-avoidant participants, but not in low fear-avoidant participants. This is in accordance with the predictions of the FA model. The findings can be interpreted as the neural correlates of anxiety induced by imagining back-straining movements. The anxiety in turn may lead to the extensive avoidance behavior shown by the high fear-avoidant patients.
Apart from the anterior hippocampus, another part of the hippocampal formation, the right parahippocampal gyrus, was also more strongly activated in high fear-avoidant than in low fear-avoidant CLBP patients. The parahippocampal gyrus has been associated with the reactivation of pain memories
The reported peak coordinates differ considerably between the studies; the coordinates resembling most closely those found in the present study were associated by Roy and colleagues
The high fear-avoidant CLBP patients also showed a stronger activation in the middle temporal gyrus than the low fear-avoidant patients. According to Palermo and colleagues,
in their activation likelihood estimation meta-analysis of fMRI studies, the middle temporal gyrus is associated with pain anticipation, perhaps indicating that HFA leads to pain anticipation when confronted with back-straining movements.
Healthy controls also revealed stronger activations in the anterior hippocampus, and in addition in the lingual gyrus, than low fear-avoidant participants. Both hippocampal clusters (differentiating the low fear-avoidant participants from the high fear-avoidant participants and from healthy controls) overlapped extensively. This was unexpected, because the healthy controls did not suffer from chronic pain and the classical FA model postulates that the movement–pain connections are learned in the acute phase of the back pain. However, even if a person does not suffer from chronic back pain, he or she may distinguish between back-straining movements and neutral movements on an affective level. After all, back-straining movements are generally more strenuous and even pain-free persons estimated them as potentially more harmful.
In our study, this is borne out by the fact that all participants, including the pain-free controls, rated the back-straining movements as more negative and arousing than the neutral ones. In healthy persons this may not influence daily living; however, when confronted with chronic pain, such fear-avoidance has a major effect on disability outcomes.
Remarkably, a direct comparison of healthy controls and high fear-avoidant patients showed no differences in neural correlates at all. The absence of a difference between healthy controls and high fear-avoidant back pain patients—even though at odds with the FA model—is in line with our first study.
In that study, the participants only passively viewed the pictures. Therefore, the question remained open as to whether the absence of differences between high fear-avoidant patients and pain-free participants could be attributed to the fact that the participants failed to relate the presented movements to their own person. In the present study, we can now rule out this explanation. In addition, Lloyd and colleagues, who compared neural responses with intense (nonpainful) tactile stimulation to the lower back between healthy controls and CLBP patients with poor adjustment did not find any differences between the 2 groups.
There are some important differences between the results of Lloyd and colleagues and the present study: They did not investigate fear-avoidance directly and used tactile lower back stimulation. However, despite these differences it is remarkable that in both studies no differences in neural responses were found between healthy controls and pain patients with low adjustment and yet both groups differed in their reaction from patients who adjust well to their chronic pain.
To analyze this unexpected similarity between high fear-avoidant participants and healthy controls further, we also showed the β differences (back-straining movements, neutral movements) for the healthy controls in the regions that differed between high fear-avoidant and low fear-avoidant patients for all 3 groups; here it also becomes apparent that for all structures for which a difference between HFA and LFA was found, the similarities between high fear-avoidant participants and controls are high, with the low fear-avoidant participants showing the diverging pattern of activation.
Instead of concentrating on the participants with HFA, perhaps we should examine more closely the participants who have experienced back pain, but report low fear avoidance. As is well known, they generally do better than persons with HFA in that they expect less pain when anticipating movements
We suspect that the low fear-avoidant participants may use some kind of coping which enables them not to react with fear to potentially painful events, but to focus on other aspects. Perhaps they concentrate on the desired outcome and the benefit of the actions and activities, rather than any pain that might accompany it. This process may be adaptive or it may represent endurance behavior as hypothesized by Hasenbring.
Pain-related avoidance versus endurance in primary care patients with subacute back pain: psychological characteristics and outcome at a 6-month follow-up.
A similar interpretation was offered by Lloyd and colleagues, who proposed that well adjusted patients use top-down processes to achieve adjustment to the pain, for instance by distracting themselves from signals arising from the lower back.
The lingual gyrus also showed stronger activation in healthy controls relative to low fear-avoidant participants. To date, little is known about the functions of the lingual gyrus, especially with respect to pain research. As part of a visuospatial neural network it codes and stores environmental information,
Regarding its role in the present context one can only speculate: Perhaps it translates into a more intensive processing of the details of the possible threatening character of the back-straining movement scenes by the healthy controls compared with the low fear-avoidant participants, who may use strategies of avoiding such processing. However, this should be addressed in future research.
Strength and Limitations
The picture stimuli were carefully selected on the basis of pilot studies with back pain patients and healthy participants. The CLBP patients in our sample were severely affected: they had been suffering from moderate to strong pain (mean pain rating of 6) for a long time (over 12 years). The high fear-avoidant patients’ average TSK score was 39 points, which is well within the range of being commonly regarded as highly fear-avoidant.
However, limitations of the study were that the fear-avoidance of the healthy sample was not assessed and that the stimulus ratings were collected retrospectively after the scanning.
Continuing the investigations of our previous study,
in which the participants may have passively viewed the back-straining movements, this time they were instructed to imagine that they themselves were carrying out the movement to engage their personal representations of the movements shown. However, although the participants have related the feared movements to themselves, the pictures may not have posed a threat to them because they knew that they would not have to perform the feared movements in reality.
Conclusions
On the basis of the FA model we expected to find fear-related neural activation in high fear-avoidant CLBP patients compared with low fear-avoidant CLBP patients and healthy controls when imagining back-straining movements. Indeed, high fear-avoidant patients showed stronger anterior hippocampus and middle temporal gyrus activation when imagining back-straining movements compared with low fear-avoidant patients, supporting the conclusion that fear-avoidance may have triggered representations related to pain expectancy to occur. However, healthy controls also had stronger hippocampus activation when contrasted with low fear-avoidant patients, contradicting an aspect of the FA model, which postulates that the movement–pain connections are learned in the acute phase of the back pain. We speculated that negative evaluations of the movement may have also been activated in healthy controls when they were confronted with back-straining movements. In contrast, the low fear-avoidant patients may have developed some kind of coping mechanism that enables them not to react with fear to back-straining movements.
Research to date has concentrated mainly on high fear-avoidant patients because they were so adversely affected. Perhaps future research should place a stronger focus on low fear-avoidant patients and investigate what may enable them to not react with fear when faced with potentially painful movements.
Acknowledgments
The authors thank the following persons for their help: Ilona Pfahlert and Britta Perl, who conducted the MRI scans, Julia Glombiewski and Jens Tersek, who helped with the data collection in the pilot study for the selection of the movement stimuli, and Elisabeth Vögtle and Johanna König, who carried out SCID I interviews.
References
Asmundson G.J.
Norton P.J.
Norton G.R.
Beyond pain: The role of fear and avoidance in chronicity.
Pain-related avoidance versus endurance in primary care patients with subacute back pain: psychological characteristics and outcome at a 6-month follow-up.
Hautzinger M: The CES-D Scale: An instrument to measure depression in investigations of the general population [in German]. Diagnostica 34:167–173, 1988
Nagel B, Gerbershagen HU, Lindena G, Pfingsten M: Development and empirical validation of the German pain questionnaire of the DGSS [in German]. Pain 16:263-270, 2002
Fear of movement and avoidance behaviour toward physical activity in chronic-fatigue syndrome and fibromyalgia: State of the art and implications for clinical practice.
30066-9&id=gr1.jpg){kind=link}
30066-9&id=gr2.jpg){kind=link}
30066-9&id=gr3.jpg){kind=link}
30066-9&id=gr4.jpg){kind=link}
30066-9&id=gr5.jpg){kind=link}