Article Figures & Data
Figures
- Figure 1.
Experimental protocol. a, Recruitment and exclusion of participants. b, Study visit timeline. c, Intratrial procedure. A CS (triangles) was presented for 2 s, participants responded with a key press to indicate which CS color was presented; 100% of CS+ were followed by a 1-s US (painful electric stimulation), each ITI trial involved a simple attention task, presenting single digits with a red cross in between, participants were asked to respond to the presentation of the red cross with a key press. CS, conditioned stimulus; US, unconditioned stimulus.
- Figure 2.
Contingency ratings, displayed as CS+/CS− differences. CS+/CS− difference in objective shock probability is 100% in acquisition and 0% in recall session. a, Placebo group, individual ratings depicted in blue. b, Doxycycline group. Individual ratings depicted in red. Mean values are marked with a horizontal black line, standard deviation is depicted with a vertical line. Time points: 1, directly after the acquisition phase; 2, acquisition contingency as remembered after the recall session; 3, after the recall session.
- Figure 3.
CS+/CS− differences in skin conductance responses (SCR) and pupil size responses (PSR) during acquisition training. a, Normalized SCR difference between CS+/CS− in the acquisition session, averaged over all trials. CS: to CS presentation, trace: during trace interval, US: to US presentation. b, Normalized PSR difference between CS+/CS− in the acquisition session, averaged over all trials. Average values are depicted with a horizontal line, standard error of mean (SEM) with a vertical line, scatterplot shows individual values, asterisk denote significant (p < 0.05) difference between CS+ and CS− trials. Detailed comparisons in Extended Data Figures 3-1 and 3-2. For sex-specific differences, see Extended Data Figures 3-3 and 3-4.
- Figure 4.
CS+/CS− differences in startle-eye blink responses (SEBR) and skin conductance responses (SCR) during recall visit 3 and their correlation with doxycycline level in serum. a, CS+/CS− differences in SEBR during recall test. Normalized SEBR difference between CS+/CS− in the recall session, averaged over all trials. b, CS+/CS− differences in SCR during recall test. Normalized SCR difference between CS+/CS− in the recall session averaged over all trials. Horizontal line: mean; vertical line: standard error of mean (SEM), asterisk denote significant (p < 0.05) difference between CS+ and CS− trials, asterisk above a line denote significant difference (p < 0.05 between placebo and doxycycline group). Individual levels of participants in the placebo group are depicted in blue, doxycycline group in red. For detailed analysis see Extended Data Figures 4-1, 4-2, 4-3, 4-4, 4-5, and 4-6. c, Correlation of CS+/CS− difference in SEBR amplitudes in doxycycline group with the doxycycline concentration in serum after acquisition session visit 2, correlation of doxycycline level and SEBR amplitudes r = −0.17. d, Correlation of CS+/CS− difference in SCR during trace interval in doxycycline group with the doxycycline concentration in serum after acquisition session visit 2, correlation of doxycycline level and SCR amplitudes r = −0.08. Individual levels of women are depicted in violet. Men are depicted in green. Dotted line shows regression regardless of sex. For more details see Table 5 and Extended Data Figure 4-7.
- Figure 5.
Correlation of body weight with doxycycline level in serum after acquisition session visit 2. Individual serum levels for women are depicted in violet. Men are depicted in green. Dotted line shows regression regardless of sex. For details, see Extended Data Figure 5-1.
- Figure 6.
Anxiety ratings. a, Placebo group, individual ratings depicted in blue. b, Doxycycline group. Individual ratings depicted in red. Mean values are marked with a horizontal black line, standard deviation (SD) is depicted with a vertical line. Asterisk above a line denote significant difference (p < 0.05) between the anxiety ratings before acquisition and after recall.
Tables
Placebo Doxycycline Sex Male (n = 24) Female (n = 24) Male (n = 24) Female (n = 25) Mean SD Mean SD p-value Cohen’s d Age (years) 24.98 4.04 24.35 3.76 0.43 0.16 Weight (kg) 71.86 11.74 69.44 13.55 0.35 0.19 US intensity (mA) 4.88 2.97 5.13 2.74 0.67 0.09 Pain ratings pre vs post 13.06 13.43 8.10 14.21 0.080 0.36 Accuracy acquisition (%) 96.93 3.54 97.35 3.97 0.58 0.11 Accuracy recall (%) 96.18 5.54 97.35 4.36 0.25 0.23 Response rate acquisition (%) 98.65 2.52 98.83 2.80 0.74 0.07 Response rate recall (%) 98.82 2.00 99.12 2.02 0.47 0.15 Differential arousal acquisition (%) 43.09 36.80 31.75 41.06 0.16 0.29 Differential arousal recall (%) 17.59 31.03 13.71 27.74 0.52 0.13 Differential valence acquisition (%) −40.68 36.06 −37.60 41.53 0.70 0.08 Differential valence recall (%) −18.30 29.30 −12.93 26.05 0.34 0.19 State anxiety preacquisition 31.65 6.37 30.53 5.46 0.37 0.19 State anxiety prerecall 31.65 6.72 28.09 4.38 0.004* 0.62 Trait anxiety preacquisition 30.94 6.53 30.29 6.88 0.64 0.10 BDI screening 3.02 3.09 2.73 2.21 0.60 0.11 BDI postrecall 2.92 4.21 3.02 3.63 0.90 0.03 US intensity: electric current used in experiment; pain ratings pre vs post: difference in average pain ratings of 14 stimuli before and after the acquisition test; accuracy: % of correct responses in identification task, average of acquisition (visit 2) and recall (visit 3); performance: % of responses in identification task, average of acquisition (visit 2) and recall (visit 3); arousal: difference in arousal ratings between CS+/CS− after the acquisition session (visit 2), and after the recall session (visit 3); valence: difference in valence ratings between CS+/CS− after the acquisition session (visit 2), and after the recall session (visit 3); state anxiety: measured with State-Trait Anxiety Inventory (STAI; Laux et al., 1981); trait anxiety: measured with State-Trait Anxiety Inventory (STAI; Laux et al., 1981); BDI: Beck Depression Inventory (Beck and Hautzinger, 2001); p-value: two-sample t test; Cohen’s d: Cohen’s d effect size of the group difference. SD: standard deviation; ‘*’ significant (p < 0.05) difference between placebo and doxycycline group.
Valence Session t df p-value Mean CS− (±SD) Mean CS+ (±SD) Placebo Acquisition 7.81 47 <0.001* 72.56 ± 21.59 31.88 ± 20.80 Recall 4.33 47 <0.001* 61.59 ± 22.49 43.29 ± 21.33 Doxycycline Acquisition 6.34 48 <0.001* 74.11 ± 21.36 36.51 ± 23.85 Recall 3.47 48 0.001* 56.56 ± 21.21 43.62 ± 20.91 Arousal Session t df p-value Mean CS− (±SD) Mean CS+ (±SD) Placebo Acquisition −8.11 47 <0.001* 21.16 ± 21.75 69.25 ± 26.99 Recall −3.93 47 <0.001* 35.73 ± 27.18 53.33 ± 26.53 Doxycycline Acquisition −5.41 48 <0.001* 36.22 ± 30.44 67.97 ± 24.83 Recall −3.46 48 0.001* 34.48 ± 29.60 48.20 ± 30.30 Valence ratings: “How do you feel when seeing this triangle? (0 = very negative, 100 = very positive)” Arousal ratings: “How aroused do you feel when seeing this triangle? (0 = very calm, 100 = very excited).” p-value: paired t-test significant differences (p < 0.05) are marked with ‘*’; SD: standard deviation.
df F value p-value Doxycycline level 1, 46 9.86 0.003* Condition (CS+/CS−) 1, 1334 0.50 0.481 Trial number 14, 1334 21.62 <0.001* Doxycycline level × Condition 1, 1334 11.54 <0.001* Doxycycline level × Trial 14,1334 1.98 0.017* Condition × Trial 14, 1334 0.61 0.856 Doxycycline level × Condition × Trial 14, 1334 2.93 <0.001* ANOVA for the relation of doxycycline levels with SEBR, dependent on condition and trial number, only for the doxycycline group. Significant effects are marked with ‘*’.
CS presentation Trace interval US time point Fear acquisition SCR F value df p-value F value df p-value F value df p-value Drug (doxycycline/placebo) 8.99 1, 1002.7 0.003* 0.43 1, 324.8 0.51 2.74 1, 266.2 0.1 Condition (CS+/CS−) 22.27 1, 3727.3 >0.001* 42.9 1, 3726 >0.001* 78.6 1, 3731.9 >0.001* Trial number 16.97 1, 3726.2 >0.001* 11.2 1, 3725.1 >0.001* 27.2 1, 3731.2 >0.001* Drug × Condition 0.61 1, 3727.3 0.435 3.68 1, 3726 0.06 3.74 1, 3731.9 0.05 Drug × Trial 11.79 1, 3726.2 >0.001* 7.23 1, 3725.1 0.007* 3.16 1, 3731.2 0.08 Condition × Trial 2.09 1, 3728.1 0.149 1.24 1, 3726.5 0.27 2.9 1, 3732.1 0.09 Drug × Condition × Trial 0.45 1, 3728.1 0.501 0.36 1, 3726.5 0.55 2.12 1, 3732.1 0.15 Fear acquisition PSR F value df p-value Drug (doxycycline/placebo) 0.2 1, 484.9 0.655 Condition (CS+/CS−) 53.5 1, 3739.4 >0.001* Trial number 0.75 1, 3738 0.387 Drug × Condition 0.63 1, 3739.4 0.429 Drug × Trial 0.06 1, 3738 0.803 Condition × Trial 3.04 1, 3739.8 0.081 Drug × Condition × Trial 1.56 1, 3739.8 0.212 Models are estimated with lme4, lmer(data ∼ (1|subject) + group*condition*trial), significant (p < 0.05) effects are marked with a * after Holm–Bonferroni correction, for robustness analysis, see Extended Data Table 3-1.
CS presentation Trace interval US time point Fear recall SCR F value df p-value F value df p-value F value df p-value Drug (doxycycline/placebo) 0.53 1, 94.19 0.468 2.13 1, 89.52 0.148 1.50 1, 1173.40 0.220 Condition (CS+/CS−) 5.55 1, 2774.23 0.019 5.57 1, 2770.09 0.018 0.56 1, 2745.20 0.455 Trial number 13.55 1, 94.11 >0.001* 38.35 1, 89.51 >0.001* 23.77 1, 2743.00 >0.001* Drug × Condition 0.02 1, 2774.23 0.889 1.66 1, 2770.09 0.198 0.10 1, 2745.20 0.749 Drug × Trial 0.72 1, 94.11 0.397 1.58 1, 89.51 0.212 1.86 1, 2743.00 0.173 Condition × Trial 2.18 1, 2774.86 0.140 0.86 1, 2771.01 0.353 0.00 1, 2746.80 0.999 Drug × Condition × Trial 0.01 1, 2774.86 0.920 0.11 1, 2771.01 0.737 0.15 1, 2746.80 0.699 Models are estimated with lme4, lmer(data ∼ (1+trial|subject) + group*condition*trial) for CS and US time point and lmer(data ∼ (1|subject) + group*condition*trial) for trace interval, significant (p < 0.05) effects are marked with a * after Holm–Bonferroni correction.
Extended Data Figure 3-1
Acquisition paired t test CS+/CS−, not corrected for multiple comparisons. Download Figure 3-1, DOC file.
Extended Data Figure 3-2
Acquisition independent t test between CS+/CS− difference for placebo and doxycycline group, not corrected for multiple comparisons. P = Placebo, D = Doxycycline Download Figure 3-2, DOC file.
Extended Data Table 3-1
SCR to CS LME (estimated with “nlme” package) in fear acquisition. Download Table 3-1, DOC file.
Extended Data Figure 4-1
Extinction paired t test CS+/CS−, not corrected for multiple comparisons. Download Figure 4-1, DOC file.
Extended Data Figure 4-2
Extinction independent t test between CS+/CS− difference for placebo and doxycycline group, not corrected for multiple comparisons. P = Placebo, D = Doxycycline Download Figure 4-2, DOC file.
Extended Data Figure 4-3
Extinction paired t test CS+/CS− per gender, not corrected for multiple comparisons. Download Figure 4-3, DOC file.
Extended Data Figure 4-4
Extinction independent t test between CS+/CS− difference for placebo and doxycycline group per gender, not corrected for multiple comparisons. P = Placebo, D = Doxycycline Download Figure 4-4, DOC file.
Extended Data Figure 4-5
SEBR ANOVA in fear recall. Download Figure 4-5, DOC file.
Extended Data Figure 4-6
SEBR LME in fear recall. Download Figure 4-6, DOC file.
Extended Data Figure 4-7
SCR and doxycycline level (in doxycycline group) LME in fear recall. Download Figure 4-7, DOC file.
Extended Data Figure 3-3
Acquisition paired t test CS+/CS− per gender, not corrected for multiple comparisons. Download Figure 3-3, DOC file.
Extended Data Figure 3-4
Acquisition independent t test between CS+/CS− difference for placebo and doxycycline group per gender, not corrected for multiple comparisons. P = Placebo, D = Doxycycline Download Figure 3-4, DOC file.
Extended Data Figure 5-1
Mediation analysis for effect of sex on doxycycline serum levels mediated by weight. Download Figure 5-1, DOC file.
In this issue
