Functional Brain Connectivity During Fear of Pain: A Comparison Between Dental Phobics and Controls

Abstract

Fear of pain is the most common reason for avoiding the dentist by patients suffering from dental phobia. Previous functional magnetic resonance imaging investigations demonstrated that already thinking about pain during the viewing of images depicting dental treatment provoked enhanced orbitofrontal cortex (OFC) activation in the clinical group. In the present study, the authors investigated whether this differential activation can be explained by differential connectivity patterns between patients and controls. They found that the control subjects displayed a stronger and more widespread connectivity compared to patients. This connectivity pattern comprised prefrontal seeds (e.g., the anterior cingulate cortex), which were coupled with limbic structures (e.g., the amygdala) and the basal ganglia (putamen, pallidum, caudate nucleus). This pattern might reflect successful emotion regulation, which was absent in the clinical group. The patients showed coupling of the OFC and the caudate nucleus, which may be the neural correlate of associating pain with dental treatment.

Introduction

Dental phobia is considered a subtype of blood-injection-injury phobia and is characterized by extreme and uncontrollable fear of dentistry. Anticipated pain and feelings of helplessness are the main reasons for the intense anxiety, which leads to avoidance of seeking dental care and subsequently to negative consequences for the oral health and overall well-being of the afflicted patients (Armfield et al., 2009; Scharmüller et al., 2014b; Schienle and Leutgeb, 2013a).

Neuroimaging research has widely neglected this phobia subtype with only a few exceptions (Lueken et al., 2011; Scharmüller et al., 2014b; Schienle et al., 2013, 2014). The first functional magnetic resonance imaging (fMRI) studies (e.g., Schienle et al., 2013) demonstrated that passive viewing of dental treatment scenes provoked increased activation within a pain modulatory network of the patients (dorsolateral/dorsomedial prefrontal cortex [DLPFC, DMPFC], orbitofrontal cortex [OFC], anterior cingulate cortex [ACC], insula, basal ganglia). More recent investigations (Scharmüller et al., 2014b; Schienle et al., 2014) focused on different attentional processes during symptom provocation. In this study, the participants were asked to direct their attention toward a foreground object within a dental treatment image (distraction), to classify the content, or to decide whether the scene elicited fear of pain. In nonphobic individuals, the last strategy was associated with increased activation of the DMPFC, which is involved in self-reference of affective meaning as well as emotion regulation (Schienle et al., 2014). Moreover, self-reported dental anxiety was positively correlated with activation in pain-processing brain regions (DMPFC, basal ganglia), when the participants focused on the pain relevance of dental treatment cues.

Scharmüller et al. (2014b) applied the same experimental design in a sample of women suffering from dental phobia and compared their neuronal responses to nonphobic controls. The clinical group only differed from the control group in the “fear of pain” condition. Both groups recruited the OFC and the amygdala with stronger OFC activation in the patients. Because the OFC and the amygdala are central for the evaluation of threat signals and negative valence assignment, their activation mirrors that the focusing on the pain relevance of the pictures was dysfunctional.

However, how was the increased OFC recruitment in the clinical group achieved? This leads to the question whether altered activation can be explained by altered connectivity in patients relative to controls. There is only one study comparing connectivity patterns between dentophobic patients and controls. Using a passive viewing task with dental treatment scenes, Scharmüller et al. (2014a) detected enhanced internal basal ganglia (putamen, pallidum) connectivity in patients. Control subjects where characterized by stronger inter-regional relationships of basal ganglia structures (putamen, pallidum) and prefrontal regions (ACC, DLPFC), which might be the basis for their successful emotion regulation. To the best of our knowledge, this fMRI study was the first one to demonstrate that dental phobia is characterized by an altered information flow within a frontostriatal network during the passive viewing of phobic stimuli.

In the present investigation, the authors reanalyzed data from Scharmüller et al. (2014b), who studied the neural correlates of different attentional strategies. As the patients had only differed from the nonphobic controls when thinking about fear of pain, the authors chose this experimental condition to explore differential connectivity patterns. They hypothesized that patients would show stronger internal basal ganglia and internal prefrontal coupling (e.g., OFC), as well as reduced interaction between the prefrontal and limbic regions. In contrast, because control subjects are able to better cope with fear of pain when watching dental treatment scenes, the authors expected that they would display stronger connectivity between the limbic regions (e.g., amygdala) and prefrontal regions involved in emotion regulation (e.g., ACC, DLPFC, DMPFC).

Materials and Methods

Sample

A total of 40 female participants were included in this investigation: 20 women who met DSM-5 (American Psychiatric Association, 2013) criteria of specific (dental) phobia and 20 nonphobic women. The two groups did not differ in age (M_patients=28.7 years, SD=9.2; M_controls=25.4 years, SD=8.0, t(38)=1.23, p=0.23) and years of education (p=0.45). All participants were right handed, medication naive, and provided their written informed consent. The study had been approved by the ethics committee of the University of Graz.

Stimuli and procedure

The participants were exposed to 28 previously validated phobic pictures showing dental treatment (Schienle et al., 2011, 2013) and 28 nonphobic images. Fourteen of the phobic pictures depicted dental treatment objects without individuals (e.g., dental driller, treatment chair), whereas the other 14 phobic pictures showed treated patients either from a first-person perspective (e.g., a dentist holds a dental driller in his hand) or from a third-person perspective (e.g., a person is treated by the dentist). The 28 nonphobic images showed objects and situations with similar physical features (e.g., color composition, complexity) as the dental treatment scenes (e.g., a chair, a fork, a person holds a pen in his hand).

The pictures were presented in three conditions, which varied in their attentional engagement: distraction, classification, and fear of pain. Because the original study (Scharmüller et al., 2014b) had observed differences between patients and controls, only for the fear of pain strategy, the reanalysis was restricted to this condition. In this study, the participants had to indicate if the picture elicited fear of pain in them. This task with an internal direction of attention explicitly required that the participants referred to their own emotional involvement (Fear of pain? Yes/no).

The pictures were presented for 2500 msec each in blocks consisting of 14 nonphobic and 14 phobic scenes in random order. Each block was initiated with the instruction (key question: e.g., fear of pain?) displayed for 3000 msec. The response options “yes”/“no” were shown on a black screen in white letters after each picture presentation. The participants used a two-button device to give the judgment for each picture. The assignment of response buttons to the answer alternatives was counterbalanced across participants. The presentation of the response options terminated when the subject pressed the button. Subsequently, a fixation cross was shown. The duration for the presentation of the response options and the fixation cross varied between 4400 and 5000 msec. Between each block, a fixation cross was displayed for 11,000 msec. The total experiment lasted ∼20 min.

After the fMRI session, the participants rated the pictures of each category using a 9-point Likert scale for valence (1=very pleasant to 9=very unpleasant) and arousal (1=not arousing to 9=very arousing). The analysis of the affective ratings showed that the symptom provocation had been successful: patients experienced more arousal (t(38)=6.6, p<0.001) and more negative valence (t(38)=9.5, p<0.001) than the control subjects when looking at dental treatment scenes. For nonphobic pictures, there were no significant group differences in affective ratings (all p's>0.58). Both groups (phobics and controls) gave lower valence ratings and higher arousal ratings for the phobic compared to nonphobic pictures (all p's<0.01).

The conducted χ ²-test showed that the patients more often experienced fear of pain when looking at the dental treatment scenes (patients: 84% vs. controls: 12%; p<0.001) than the controls, whereas there was no difference for the nonphobic pictures (patients: 0.3% vs. controls: 0.1%). Full details of the procedure and affective ratings have been published previously (Scharmüller et al., 2014b).

Functional magnetic resonance imaging

The fMRI session was conducted with a 3T scanner (Skyra; Siemens, Erlangen, Germany). For functional runs, a total of 494 volumes were acquired using an echo-planar imaging protocol (number of slices: 35, descending, tilted −25° from the AC-PC line; flip angle=90°; slice thickness: 3 mm; matrix: 64×64; TE=30 msec; TR=2290 msec; FoV: 192; in-plane resolution=3×3 mm). All analyses were conducted using SPM8 (Wellcome Department of Cognitive Neurology, London). Three volumes from the beginning of the time series were discarded to account for saturation effects. Functional data were realigned, including unwarping, normalized to Montreal Neurological Institute space (3-mm isotropic voxel), and smoothed with an 8-mm isotropic Gaussian kernel. Individual conditions were modeled using the canonical hemodynamic response function. Data were high pass filtered (128 sec). Temporal sphericity was controlled by an AR(1) process with consecutive prewhitening of the data.

Psychophysiological interaction analysis

The authors investigated phobia-related functional connectivity for each participant within the “fear of pain” condition using psychophysiological interaction (PPI) analyses (Friston et al., 1997). In terms of condition-specific covariation in residuals, PPIs assess the extent to which an experimental factor modulates the connectivity of one brain region with others. The PPI identifies voxels that covary differentially with a given seed region (e.g., left/right ACC) as a function of an experimental factor. For each subject, the individual PPI design matrix consists of three regressors: the time series of activity taken from the seed region (=physiological variable), the condition type (e.g., the contrast Phobia>Neutral;=psychological variable), and as the third regressor, the PPI variable, which comprises the calculation of an interaction term by an element-by-element product between the first and second regressor. To compare connectivity for the contrast: Phobia>Neutral between phobic and control subjects, the subject-specific interaction contrasts were entered into a random-effects analysis (thresholded at p<0.05, corrected for multiple comparisons [familywise error]).

The authors defined 14 seed regions (caudate nucleus, pallidum, putamen, DLPFC, ACC, amygdala, OFC; each region for the right and the left hemisphere) based on previous findings on disorder-specific activation in dental phobia (Scharmüller et al., 2014b; Schienle et al., 2013). The same regions were defined as regions of interest (ROIs), which were taken from the automatic anatomical labeling atlas (Tzourio-Mazoyer et al., 2002) and from the Harvard-Oxford cortical and subcortical structural atlas (Mazziotta et al., 2001).

Results

In the “fear of pain” condition (for the contrast: Phobia>Neutral), the patient group showed functional coupling between the frontal regions (DMPFC as seed with the ACC and the DLPFC) as well as coupling of the OFC with the caudate nucleus.

Nonphobic participants showed internal connectivity of frontal areas (ACC, DMPFC, OFC) as well as coupling of limbic structures (amygdala) with frontal regions (ACC, DLPFC, DMPFC). For all results, see Table 1.

Table 1.

Functional Connectivity in the “Fear of Pain Condition” for the Contrast Phobia>Neutral

Seed	Region	H	X	Y	Z	T	P(FWE)	Cluster size
Within Phobics
DMPFC left	ACC	L	−9	23	25	4.42	0.030	25
	DLPFC	R	18	14	49	5.24	0.036	880
OFC left	Caudate	L	−15	14	7	4.83	0.008	108
	Caudate	R	15	2	19	3.91	0.045	190
Within Controls
ACC left	ACC	L	−3	23	19	5.10	0.009	80
	ACC	R	6	29	19	4.49	0.024	146
	DMPFC	L	−6	20	40	5.07	0.017	103
Amygdala left	ACC	R	6	11	25	5.24	0.006	171
Amygdala right	ACC	L/R	0	26	22	5.03	0.011	369
	ACC	R	6	11	25	5.14	0.008	300
	Pallidum	L	−9	−1	−5	3.62	0.045	29
DLPFC left	ACC	L	−3	32	19	5.42	0.005	186
	ACC	R	6	32	16	5.29	0.005	169
DLPFC right	ACC	L/R	0	29	16	4.99	0.010	314
	ACC	R	6	29	19	5.90	0.001	274
DMPFC left	ACC	L	−6	23	22	6.83	<0.001	218
	ACC	R	6	32	16	8.05	<0.001	291
	Caudate	R	18	26	−2	4.28	0.025	146
	DMPFC	L	−9	20	40	6.37	0.001	39
DMPFC right	ACC	L	−6	23	22	5.63	0.003	263
	ACC	R	6	32	16	6.39	<0.001	304
OFC left	ACC	L	−3	26	22	5.02	0.009	311
	ACC	R	6	29	19	4.86	0.011	292
	OFC	R	3	32	−11	5.37	0.016	385
OFC right	ACC	L/R	0	29	19	4.98	0.010	238
	ACC	R	6	29	16	5.82	0.002	231
Phobics>Controls
DLPFC right	Pallidum	R	18	−1	4	3.33	0.036	11

Controls>Phobics
Amygdala left	ACC	R	6	11	25	3.98	0.023	385
	Caudate	R	12	17	−5	3.69	0.036	209
	Pallidum	R	15	−1	1	3.49	0.026	73
	Putamen	R	18	17	−8	3.69	0.045	249
Amygdala right	ACC	L/R	0	41	−5	4.15	0.015	320
	OFC	R	3	41	−5	4.36	0.027	216
OFC left	ACC	L/R	0	35	−5	3.66	0.049	206
	ACC	R	3	35	−5	3.93	0.022	139
	DLPFC	L	−51	29	31	4.82	0.012	153
OFC right	ACC	L/R	0	38	−2	3.88	0.028	171
	ACC	R	3	35	−2	3.82	0.028	152
Pallidum right	ACC	R	9	29	22	3.73	0.039	282
Putamen right	ACC	L	−6	38	1	4.50	0.007	334
	ACC	R	3	35	−2	4.01	0.021	309
	Caudate	L	−18	8	16	3.63	0.041	69

ACC, anterior cingulate cortex; DLPFC, dorsolateral prefrontal cortex; DMPFC, dorsomedial prefrontal cortex; H, hemisphere; OFC, orbitofrontal cortex; P(FWE), p-value, familywise error corrected; seed, seed region for psychophysiological interaction analysis; T, t-value; X,Y,Z, MNI (Montreal Neurological Institute) coordinates.

The authors also compared connectivity patterns between groups. Results indicated that nonphobic subjects where characterized by a widespread connectivity network. Relative to the clinical group, they showed stronger functional coupling of the amygdala (seed) with the ACC, OFC, and the basal ganglia (caudate nucleus, putamen, pallidum). The OFC (seed) showed enhanced connectivity with the ACC and the DLPFC. The basal ganglia regions (putamen, pallidum) showed stronger coupling with the ACC.

In contrast, patients compared to nonphobic controls showed enhanced functional connectivity only for the DLPFC, as seed with the pallidum (see Fig. 1 and Table 1).

FIG. 1.

Group differences in functional connectivity. DLPFC, dorsolateral prefrontal cortex; OFC, orbitofrontal cortex. Color images available online at www.liebertpub.com/brain

Discussion

The present PPI analysis showed differential connectivity patterns between dentophobic patients and controls while they were asked to direct their attention to the pain relevance of images depicting dental treatment.

Within the patient group, functional connectivity was only present between the DMPFC (seed) and the ACC, as well as the DLPFC. Furthermore, the OFC (seed) was coupled with the caudate nucleus. OFC activation can be found in normal as well as phobic individuals (Milad and Rauch, 2007) and has a central role in affective learning. The functional coupling of the OFC with the basal ganglia (caudate nucleus) may have promoted the learning that dental treatments are painful. This may be the case because the basal ganglia are involved in a pain modulatory network and integrate many aspects of pain (e.g., emotional, autonomic, cognitive, or motor responses; Seger and Cincotta, 2005). The connectivity between these two areas while focusing on the pain relevance of a stimulus possibly enhanced the process of associating pain with dental treatments.

For controls, the authors found internal functional coupling within prefrontal areas (ACC, OFC, DLPFC) as well as coupling of limbic regions (e.g., amygdala as seed) with prefrontal regions (ACC, OFC). Additionally, the authors observed connectivity of the amygdala (seed) and the basal ganglia (caudate nucleus, putamen, pallidum) and of the basal ganglia (seed) with the ACC. It is of interest that the detected network was widespread in the control group. This finding indicates that the nonphobic individuals were characterized by more complex information processing than the patients and may therefore be able to better cope with the situation. Numerous fMRI investigations identified (co)activation in frontolimbic circuits as crucial for successful downregulation of negative affective states (for a meta-analysis, see Buhle et al., 2014). Moreover, Gold et al. (2014) showed that the exposure to unpredictable threat modulated amygdala–prefrontal cortex connectivity in healthy subjects, which helped to maintain adequate performance.

A similar differential pattern has also been described by Postuma and Dagher (2005). The authors showed that healthy individuals were characterized by a more widespread coactivation of striatal nuclei and prefrontal regions like the ACC and the DLPFC. This extended coactivation pattern was also prominent in the current analyses for nonphobic individuals. In patients, they observed reduced functional coupling indicating a more rigid processing style concerning phobic stimuli, which possibly leads to more fear of pain and to stronger subjective suffering.

Taking these findings together, focusing on the pain relevance in a dental treatment situation seems to be a dysfunctional strategy to reduce anxiety (Scharmüller et al. 2014b). While healthy individuals automatically recruit emotion regulation areas to control limbic regions, this is not the case in patients (Buhle et al., 2014). The stronger OFC activation (Scharmüller et al., 2014b) and the enhanced internal frontal coupling found in patients underline the negative valence assignment in patients compared to controls. As Rolls (2004) stated, the OFC has an important role for establishing and controlling reward-related and punishment-related behavior and therefore modulates emotions. The lack of appropriate connectivity of limbic structures and frontal regions in patients may lead to a deficient emotional processing and therefore to a stronger recruitment of the OFC, which is the first cortical region where the subjective (un)pleasantness of stimuli is made explicit (Rolls, 2013).

These findings have important clinical implications. The commonly recommended coping strategy by dentists for their patients “please raise your hand when you start to feel pain” seems to be not helpful at least for anxious or phobic individuals because it directs the attention toward possible pain. Obviously, alternative attentional strategies are needed, which help the patient to attend to nonthreatening features of the phobic stimulus. Interestingly, the instruction to classify dental treatment scenes did not reveal activation differences between dental phobics and controls (Scharmüller et al., 2014b). This indicates that this strategy reduced processing differences between the two groups. Future studies are needed to investigate whether an objective focusing on disorder-relevant stimulus features might even help to change the rigid central pain modulatory networks of dentophobic patients. In this sense, connectivity analyses could be used to evaluate the effects of attentional trainings.

The authors have to mention the following limitations of the present investigation. They interpreted their data as a reflection of differential processing of fear of pain in dentophobic patients and controls. However, it has to be kept in mind that the observed connectivity patterns do not only reflect “fear of pain”, but are the result of a mixture of different cognitive-affective processes. The task required the participants to keep the instruction in working memory, to conduct self-reflection, and to prepare for decision-making. These overlapping, and perhaps even interfering processes, cannot be separated by means of connectivity analysis. Additionally, the authors could not investigate whether these results are specific for dental phobia as the study design lacks another clinical group. This should be done in future studies. Finally, they only investigated female patients and controls. Therefore, the results cannot be generalized to men.

Conclusion

This reanalysis with a PPI approach showed that differences in localized activation (stronger OFC activation in patients afflicted with dental phobia relative to nonphobic controls during visual symptom provocation) were accompanied by differences in functional integration in frontostriatal circuits.

Footnotes

Author Disclosure Statement

The authors declare that they have no conflict of interests.

References

American Psychiatric Association.. 2013. Diagnostic and Statistical Manual of Mental Disorders, 5th ed. Arlington, VA: American Psychiatric Publishing.

Armfield

, Slade

, Spencer

. 2009. Dental fear and adult oral health in Australia. Community Dent Oral Epidemiol, 37:220–230.

Buhle

, Silvers

, Wager

, Lopez

, Onyemekwu

, Kober

, Weber

, Ochsner

. 2014. Cognitive reappraisal of emotion: a meta-analysis of human neuroimaging studies. Cereb Cortex, 24:2981–2990.

Friston

, Buechel

, Fink

, Morris

, Rolls

, Dolan

. 1997. Psychophysiological and modulatory interactions in neuroimaging. Neuroimage, 6:218–229.

Gold

, Morey

, McCarthy

. 2014. Amygdala-prefrontal cortex functional connectivity during threat-induced anxiety and goal distraction. Biol Psychiatry pii: S0006-3223(14)00265-0.

Lueken

, Kruschwitz

, Muehlhan

, Siegert

, Hoyer

, Wittchen

H-U

. 2011. How specific is specific phobia? Different neural response patterns in two subtypes of specific phobia. Neuroimage, 56:363–372.

Mazziotta

, Toga

, Evans

, Fox

, Lancaster

, Zilles

, Woods

, Paus

, Simpson

, Pike

, Holmes

, Collins

, Thompson

, MacDonald

, Iacoboni

, Schormann

, Amunts

, Palomero-Gallagher

, Geyer

, Parsons

, Narr

, Kabani

, Le Goualher

, Boomsma

, Cannon

, Kawashima

, Mazoyer

. 2001. A probabilistic atlas and reference system for the human brain: International Consortium for Brain Mapping (ICBM). Philos Trans R Soc Lond B Biol Sci, 356:1293–1322.

Milad

, Rauch

. 2007. The role of the orbitofrontal cortex in anxiety disorders. In: Schoenbaum

, Gottfried

, Murray

, Rasmus

(eds.) Linking Affect to Action: Critical Contributions of the Orbitofrontal Cortex. Oxford: Blackwell Publishing, pp. 546–561.

Postuma

, Dagher

. 2005. Basal ganglia functional connectivity based on a meta-analysis of 126 positron emission tomography and functional magnetic resonance imaging publications. Cereb Cortex, 16:1508–1521.

10.

Rolls

. 2004. The functions of the orbitofrontal cortex. Brain, 55:11–29.

11.

Rolls

. 2013. A biased activation theory of the cognitive and attentional modulation of emotion. Front Hum Neurosci, 18:7–74.

12.

Scharmüller

, Leutgeb

, Schienle

. 2014a. Altered functional connectivity of basal ganglia circuitry in dental phobia. Soc Cogn Affect Neurosci, 9:1584–1588.

13.

Scharmüller

, Übel

, Leutgeb

, Schöngassner

, Wabnegger

, Schienle

. 2014b. Do not think about pain: neural correlates of attention guiding during visual symptom provocation in dental phobia—an fMRI study. Brain Res, 1566:69–76.

14.

Schienle

, Köchel

, Leutgeb

. 2011. Frontal late positivity in dental phobia: a study on sex differences. Biol Psychol, 88:263–269.

15.

Schienle

, Leutgeb

. 2013. Differential neurobiology of affective processing in specific phobia—implications for psychotherapy. Z Psych Psychol Psychother, 61:105–112.

16.

Schienle

, Scharmüller

, Leutgeb

, Schäfer

, Stark

. 2013. Sex-specific brain activation and structure in dental phobia. Brain Struct Funct, 218:779–787.

17.

Schienle

, Wabnegger

, Schoengassner

, Scharmüller

. 2014. Neuronal correlates of three attentional strategies during affective picture processing: an fMRI study. Cogn Affect Behav Neurosci, 14:1320–1326.

18.

Seger

, Cinotta

. 2005. Dynamics of frontal, striatal, and hippocampal systems during rule learning. Cereb Cortex, 16:1546–1555.

19.

Tzourio-Mazoyer

, Landeau

, Papathanassiou

, Crivello

, Etard

, Delcroix

, et al. 2002. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage, 15:273–289.