More than fear? Brain activation patterns of dental phobic patients before and after an exposure-based treatment

Seventeen patients diagnosed with dental phobia (9 female/8 male) and 17 non-phobic controls (10 female/7 male) participated in this study, see Table 1. All phobic individuals were recruited at the ‘Treatment Centre for Dental Fear’ at a local dental clinic in Bochum (Germany) where they sought psychological care due to high and impairing dental fear. Individuals were asked to participate in the study if they met the criteria for a Specific (Dental-) Phobia according to the standards of DSM-IV (American Psychiatric Association 2000), did not take any psychopharmacological drugs, and were not pregnant, pierced or tattooed. Diagnoses and medical exclusion criteria were confirmed using the short form of the German Mini-DIPS (Margraf 1994) as a semi-structured clinical interview. The following comorbid disorders were diagnosed in DP-patients: 3 Specific Phobias (in addition to DP), 1 Social Phobia, 3 Substance Withdrawals (excluding nicotine).

Control subjects were asked to participate if they were not dental phobic, i.e. did not report any symptoms of dental fear or avoidance of dental treatments, scored beyond the cut-off score for mild dental fear in the Dental Anxiety Scale, i.e. displaying a score ≤ 12 (scale description see below) and denied any form of psychological treatment or psychotropic drug taking. Since no clinical interview was conducted with the control subjects, no statement can be made about any existing mental disorders in this subsample besides dental phobia.

All participants gave written informed consent before functional imaging started. The local ethics committee of the psychological faculty of the Ruhr University Bochum approved the study (number 060).

A stress inoculation training (Meichenbaum 2007) adapted for the use in DP patients (Sartory and Wannemüller 2010; Wannemüller et al. 2015b) was administered in five standardized sessions imparting cognitive and bodily coping strategies to apply them in various exposure exercises. Session 1 consisted of conducting the diagnostic interview and imparting psycho-educative information about the sense of fear and its evolutionary function. Patients’ cognitive and physiological symptoms in a dental surgery situation were gathered and the rationale of helpful thoughts, breathing techniques and applied relaxation as cognitive and bodily coping strategies was explained. In order to prepare for the later usage in the dental situation, patients were asked to train Progressive Muscle Relaxation (PMR) in home exercises via CD-instructions. In the second session, a three-step program for applied relaxation (Öst 1987) based on the PMR concept was introduced. The program aimed at teaching an individualized form of short relaxation to be trained at home and applied against bodily fear symptoms in highly stressful situations. Moreover, the patients elaborated coping-oriented thoughts for the dental situation with the help of the therapist. The third session started with a video exposure of a dental situation (filling of a carious tooth). During the scene, patients should concentrate on their cognitive and physiological fear responses. Afterward, patients were invited to undergo a noise exposure exercise consisting of different dental-burr noises (turbine burr and rose-head burr). During exposure, patients were asked to cope with eliciting fear by applying a deep abdominal breathing technique instructed by the psychotherapist. The fourth and fifth sessions both consisted of therapist-guided in sensu exposure exercises both composed of vividly descriptions of a dental surgery. In the fourth session, the use of newly learned bodily and cognitive coping strategies was prompted by the therapist. In the fifth session, patients should cope with upcoming fear responses during exposure. Before treatment ended, patients required relapse-prevention strategies and were encouraged to directly arrange a dental appointment. Against the background of the treatment components described and the rationale of stress inoculation training, active coping of anxiety symptoms and the associated increase in the experience of control is the postulated key mechanism of action of the applied treatment. However, since repeated exposure exercises were also carried out with the patients, it is likely that habituation and inhibitory learning processes may also have occurred during treatment.

The fMRI experiment was conducted at a local Medical Centre in Bochum, Germany (Grönemeyer Institute for Microtherapy). The pre-treatment scan of patients was conducted in the interval between treatment session 1 and treatment session 2, i.e. after the diagnostics and psychoeducation session and prior to the experience of any exposure treatment elements. Post-treatment scan was conducted within two weeks after completing the treatment.

The 20 neutral pictures all derived from the International Affective Picture System (IAPS) (Lang et al. 2001). Neutral sounds were picked from the International Affective Digitalized Sound System (IADS) (Bradley and Lang 1999). Dental-related pictures were partly taken from free web-sources and partly from the IAPS. Dental related sounds were self-recorded and consisted of 12 dental burr noises (6 turbine burrs; 5 rose-head burrs), three ultrasonic tartar removers, three sonic tartar removers, and three suction devices to remove saliva from the mouth, see Table S1 for an overview of the applied stimuli.

In order to assess dental fear on different levels we applied several questionnaires. We used the German version of the Dental Anxiety Scale (DAS) (Corah 1969) and the German ‘Hierarchischer Angstfragebogen’ (HAF) (Johren 1999) to assess the subjective component of dental fear. In both questionnaires, participants are asked to rate the extent of subjective fear in four (DAS) or respectively eleven (HAF) dental-related situations on a five-item scale (1–5). The authors of both instruments report sufficient to good internal consistencies, ranging from Cronbachs’s α = 0.64 (DAS) to α = 0.80 (HAF). Scores ≥ 15 in the DAS are considered to indicate high dental fear and scores of 13 and 14 moderate dental fear (Corah et al. 1987). According to the author of the HAF, a score > 38 very sensitively indicates dental phobia.

A total of 440 T2*-weighted whole brain volumes were acquired using a 1.5 T Symphony scanner (Siemens, Germany). Each volume consisted of 25 slices of 3 mm with an inter-slice gap of 1 mm. The repetition time was 80 ms per slice with a flip angle of 90°. We further acquired a high-resolution T1-weighted image with a voxel size of 1 × 1 × 1 mm and a repetition time of 2110 ms. The echo time was 3.93 ms with a flip angle of 15°. T1-weighted images were used for neuroanatomical localization and co-registration of the functional data.

Functional data were analyzed using the latest version of SPM 12 software package (https://​www.​fil.​ion.​ucl.​ac.​uk/​spm/​). Images were slice time corrected realigned and unwarped in a second step. In the next step, these data were segmented using the tissue probability maps provided by SPM12. Images were then co-registered to the warped mean image. In the last step, images were normalized and smoothed with a Gaussian kernel of 8 mm. All functional imaging data were fed into a first-level analysis including five regressors: visual neutral images, visual phobic images, auditory neutral stimuli, auditory phobic stimuli and rating interval. Phobic stimuli were contrasted with neutral stimuli of the same modality. These contrast images were fed into a second-level analysis using an ANOVA with the factors Modality (visual and auditory) and Group (patients and healthy controls). Main effects and interactions were analysed using non-directional f-contrasts. Data were thresholded at p < 0.05 using a false discovery rate (fdr) correction and a minimum of 8 contiguous voxel. The resulting activation clusters were defined as regions of interest (ROI) for a pre-post analysis of patients’ activation patterns. Signal changes of these ROIs were extracted using the marsbar region of interest toolbox 0.44 (http://​marsbar.​sourceforge.​net/​). Extracted signal changes were further analysed using the statistical software package SPSS (IBM SPSS 26) by applying an ANOVA with the factors Modality (visual and auditory) and Time (pre and post) to all ROIs separately.

Sample characteristics, clinical data, subjective stimulus ratings as well as the results of pre- to post-treatment comparisons and between-subject comparisons are presented in Table 1. As intended, patients with DP did not differ from controls in regard to age, sex or educational level and showed highly significantly larger scores in both instruments measuring dental fear. This still–however to a much less extent–was the case when comparing the post-treatment scores of phobic participants to that of controls. Nevertheless, within phobic participants we observed substantial improvements across all dental instruments from pre- to post-treatment dental fear (mean change = 32.38% ± 17.78%).

Concerning the stimulus ratings at pre-treatment assessment a 2 (Group: DP vs. HC) × 2 (Modality: auditory vs. visual) × 2 (Valence: phobia related vs. neutral) repeated measure ANOVA yielded highly significant main effects of Group, F(1,28) = 35.14, p < 0.001, η² = 0.58 (DP patients > HC); Modality, F(1,28) = 5.60, p = 0.025, η² = 0.17 (sounds > pictures) and Valence, F(1,28) = 129.51, p < 0.001, η² = 0.82 (dental > neutral). The Group x Valence effect, F(1,30) = 36.17, p < 0.001, η² = 0.56, was the only significant interaction effect, see Table 1 for post-hoc test results. Repeated measures ANOVA within DP-patients yielded a highly significant effect of Time, F(1,13) = 14.48, p = 0.002, η² = 0.53 (pre > post) and Valence, F(1,15) = 71.92, p < 0.001, η² = 0.85 (dental > neutral). No interaction effect was significant.

When comparing the post-treatment stimulus ratings of DP patients with that of HCs the Group x Valence effect no longer existed, F(1,31) = 2.86, p < 0.03, η² = 0.13, see Table 1 for post-hoc test results.

No voxel survived the threshold for the interaction between Group and Modality. Only the main effect for Group yielded evidence for activation differences at the defined threshold of p < 0.05 (fdr corrected) in a big cluster covering the right precuneus and extending into the lateral parietal cortex covering Brodmann area (BA) 39 and BA 40 (see Table 2). Additionally, a cluster in the midline near precuneus (BA 7) was found. Further clusters were found in the left primary and right premotor cortex. The right supramarginal gyrus was also activated in this main effect. Further activation foci were found in the right insula, the anterior cingulate cortex (ACC) as well as in the superior temporal gyrus, see Table 2 and Fig. 1.

The former described contrast illustrates the general activation difference between healthy controls and DPs during processing of visual and auditory phobic stimuli. To investigate pre-post related treatment changes and taking the general differences between healthy controls and DPs into account, activation clusters from the former described main effect (Group: healthy controls vs DP-patients pre-treatment) were used as ROIs for further analysis. Signal changes were extracted from these ROIs for pre- and post-treatment activation and fed into a repeated measures ANOVA including the factors Modality, Valence and Time (pre and post-treatment) in search for Valence × Time effects or Modality × Valence × Time effects. Results from all these analyses yielded evidence for significant Valence × Time effects in all extracted ROIs. Again, in no case any interaction effects containing the factor Modality or triple interaction effects reached the level of significance, see Fig. 2 for the course of signal changes in the respective ROIs.

Correlation analyses examining associations between dental fear and signal changes at pre-treatment assessment in the extracted ROIs yielded a significant correlation between the signal change to phobia-related material in the ACC (BA 32) and pre-treatment subjective dental fear (see Fig. 3a).

Moreover, pre- to post-treatment activation decrease in this area was correlated with changes in dental fear after treatment (see Fig. 3b).