SUPPLEMENTARY MATERIALS

Birth control

The acceptable methods of birth control included: a double-barrier method (condoms or diaphragm) combined with spermicide; hormonal contraceptives plus single barrier (birth control pills, implants [Norplant] or injections [Depo-Provera]), Intrauterine Device (IUD), or abstinence (no sexual activity).

Effect of oral contraception on baseline startle and fear-potentiated startle

To examine the potential effect of oral contraception variation in hormone levels during the menstrual cycle on baseline startle and fear-potentiated startle, we examine pre- and post-drug baseline startle reactivity in the placebo condition and fear-potentiated startle during placebo and high GSK in participants who were taking or were not taking oral contraception (see Table below).

The table below shows baseline startle reactivity and fear-potentiated startle in the placebo and high GSK condition in 24 participants who did not use oral contraception and eight participants who used oral contraception. We do not present statistical comparison between groups as the sample size of the oral contraception group is too low to detect small effect sizes. However, inspection of the data suggests that baseline startle was not affected by oral contraception. The result of increased fear-potentiated startle in the high GSK condition seems to hold for both groups.

Baseline startle pre-drug (µVolts) / Baseline startle post-drug
(µVolts) / Anxiety-potentiated startle
(difference t-scores)
Plc / Plc / Plc / High GSK
Oral contraception (n=8) / 85.3 (12.8) / 87.5 (12.4) / 5.2 (2.1) / 10.5 (2.9)
No oral contraception (N=23) / 99.1 (27.9) / 85.5 (21.5) / 7.8 (1.3) / 11.3 (1.7)

Startle habituation during threat blocks 1 and 2

The Table below presents startle magnitude (raw scores) during the ITI N condition of threat blocks 1 and 2. It shows that startle magnitude was similarly reduced with the alprazolam and high GSK treatments in block 1 and block 2. A 2-way ANOVA Treatment (4) x Threat Block (1,2) on the ITI startle magnitude shows main Treatment (F(1,30)=4.7, p<.01) and Block (F(1,30)=20.0, p<.001) effects but no Treatment x Block interaction (F(3,90)=.8, ns (same results with t-scores). These results suggest that the drug reduced startle reactivity and that reduction did not change over time.

Plc / Alp / Low GSK / High GSK
Threat block 1 / 58.9 / 47.5 / 53.7 / 45.4
Threat block 2 / 48.1 / 30.1 / 41.9 / 32.0