Aerobic exercise reduces anxiety and fear ratings to threat and increases circulating endocannabinoids in women with and without PTSD

Abstract

Reductions in state anxiety have been reported following an acute bout of aerobic exercise. However, less is known regarding anxiety and fear ratings to specific threatening stimuli following an acute bout of aerobic exercise in women with PTSD. Moreover, the mechanisms responsible for the anxiolytic effects of exercise are not fully understood, although recent studies suggest a role for the endocannabinoid (eCB) system. Thus, this study utilized a randomized, counterbalanced approach to examine anxiety and fear ratings to predictable or unpredictable electric shock administration and circulating concentrations of eCBs and mood states immediately following moderate-intensity aerobic exercise (30 min on treadmill at 70–75% maximum heart rate) and a quiet rest control condition in women with and without a history of trauma, and in women with PTSD (N=42). Results revealed that anxiety and fear ratings to unpredictable and predictable threats were significantly (p<.05) lower following exercise compared to quiet rest, with correlational analyses indicating those with greater increases in circulating eCBs had greater reductions in anxiety and fear ratings to unpredictable and predictable threats following exercise. Also, there were significant (p<.05) reductions in fatigue, confusion, total mood disturbance, and increases in positive affect following exercise for the entire sample. Non-trauma controls and PTSD groups reported significant (p<.05) increases in vigor, with additional mood improvements following exercise for the PTSD group (i.e., decreases in state anxiety, negative affect, tension, anger, and depression). Results from this study suggest that aerobic exercise exerts psychological benefits in women with PTSD, potentially due to exercise-induced increases in circulating concentrations of eCBs.

Introduction

Although the anxiolytic effects of acute aerobic exercise have been well-established, the mechanisms underlying this effect are not fully understood (Bahrke & Morgan, 1978; Garvin, Koltyn, & Morgan, 1997; O’Connor, 2005; Herring, Monroe, Gordon, Hallgren, & Campbell, 2019). Moreover, the majority of supporting evidence for an anxiolytic effect of aerobic exercise has been derived from investigations primarily focused on assessing self-reported state anxiety prior to and following participation in an acute bout of exercise (Bahrke & Morgan, 1978; Garvin, Koltyn, & Morgan, 1997; O’Connor, 2005; Herring, Monroe, Gordon, Hallgren, & Campbell, 2019). While these investigations have been informative, there has been a scarcity of research examining the effect of aerobic exercise on anxiety and fear responses to specific stimuli capable of inducing anxiety and fear and manipulating participant’s affect (e.g., predictable and unpredictable shock during a lab-based task), which would inform us of the potential influence of aerobic exercise on these outcomes in response to experiencing specific threats, as opposed to more general state anxiety responses following exercise.

To date, only one prior study has administered a lab-based task and examined anxiety and fear responses to unpredictable and predictable threat following aerobic exercise. Specifically, Lago and colleagues (2018) found that an acute bout of moderate-intensity cycling (compared to a light-intensity control condition) significantly reduced startle potentiation during an unpredictable threat condition but not during a predictable condition in 34 healthy male and female young adults. Given that anxiety is thought to involve a sustained state of apprehension evoked by an unpredictable threat, while fear involves a more phasic response to a predictable threat (Davis, Walker, Miles, & Grillon, 2010) the authors concluded that exercise was able to attenuate anxiety responses to threat (specifically anxiety potentiated startle; Lago et al., 2018). However, given that only one study has been conducted, there is a need to expand these findings to clinical populations where anxiety and fear are typically elevated (e.g., Posttraumatic Stress Disorder, PTSD), in addition to examining potential mechanisms responsible for these effects.

PTSD is a highly debilitating disorder that can develop following exposure to one or more potentially traumatic events (e.g., interpersonal violence, motor-vehicle accident, death of a close loved one, serving in combat zone; American Psychological Association, 2013). In addition to PTSD specific symptoms (avoidance behavior, re-experiencing, negative alterations in cognition, mood, arousal, and reactivity), adults with PTSD typically experience numerous anxiety symptoms and are often diagnosed with a comorbid anxiety disorder (American Psychological Association, 2013; Brown, Campbell, Lehman, Grishman, & Mancill, 2011; Kessler et al., 2005). Only recently have the acute anxiolytic effects of aerobic exercise been examined in this population, with reports suggesting that moderate-intensity aerobic exercise results in significant reductions in self-reported state anxiety symptoms (Crombie, Brellenthin, Hillard, & Koltyn, 2018; Crombie, Leitzelar, Brellenthin, Hillard, & Koltyn, 2019). Recent research suggests that the reported reductions in state anxiety and overall improved mood may be due to the influence of aerobic exercise on an expansive neuromodulatory system known as the endocannabinoid (eCB) system (Crombie et al., 2018; Crombie et al., 2019; Raichlen, Foster, Gerdeman, Seillier, & Giuffrida, 2012; Brellenthin, Crombie, Hillard, & Koltyn, 2017).

The eCB system, which is widely distributed throughout the central and peripheral nervous systems, primarily consists of receptors (cannabinoid type-1, CB1R; cannabinoid type-2, CB2R) and endogenous ligands (eCBs; N-arachidonoylethanolamine [AEA]; and 2arachidonoglycerol [2-AG]; Matsuda, Lolait, Brownstein, Young, & Bonner, 1990; Herkenham, Lynn, Johnson, Melvin, de Costa, & Rice, 1991). Given that eCB/CB1R signaling modulates synaptic transmission throughout the limbic circuit, the eCB system has emerged as playing an important modulatory role in regulating numerous neuronal functions, including anxiety, fear, and stress processes and behavioral outputs (Lafenetre, Chaouloff, & Mariscano, 2007; Katona & Freund, 2012; Lutz, Mariscano, Maldonado, & Hillard, 2015). Moreover, previous research has reported on significant relationships between aerobic exercise-induced increases in circulating concentrations of eCBs and improvements in positive mood states, and reductions in both negative mood states and self-reported state anxiety (Crombie et al., 2018; Crombie et al., 2019; Raichlen et al., 2012; Brellenthin et al., 2017; Meyer, Crombie, Cook, Hillard, & Koltyn, 2019). However, the role that the eCB system plays in reducing anxiety and fear responses specifically to unpredictable and predictable threats remains unknown.

Therefore, the current study sought to further our understanding of anxiety and fear responses to unpredictable and predictable threats following an acute bout of aerobic exercise by examining a neurobiological mechanism potentially responsible for the hypothesized reductions in adult women with and without a clinical psychiatric disorder. Specifically, the current study utilized a randomized, counterbalanced approach to examine self-reported anxiety and fear ratings to unpredictable and predictable threats and circulating concentrations of eCBs following exercise and quiet rest in 3 groups of adult women: those not been exposed to trauma (i.e., trauma free), and adult women exposed to trauma with and without a current diagnosis of PTSD. The current study also assessed self-reported state anxiety, mood states, and eCBs prior to and following aerobic exercise and quiet rest. It was hypothesized that 1) circulating concentrations of eCBs would significantly increase following moderate-intensity aerobic exercise but not quiet rest (for all three groups), 2) self-reported anxiety and fear ratings to the unpredictable and predictable threats administered during the task would be attenuated (for all three groups) following exercise compared to quiet rest, and 3) greater differences in eCB responses between conditions (i.e., change in eCBs following aerobic exercise – change in eCBs following quiet rest) would be positively associated with differences in fear and anxiety responses between conditions (i.e., fear and anxiety responses during quiet rest – fear and anxiety responses during exercise).

Methods

All procedures were approved by the Health Sciences Institutional Review Board at the University of Wisconsin-Madison and the work described has been carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans.

Study Participants

A power analysis (G*Power 3.1) was conducted to determine the sample size needed per group to detect a significant increase in eCBs following exercise using a repeated measures, within-between interaction design. The analysis (which was based on a previously published report examining eCB responses in adults with PTSD; Crombie et al., 2018) was powered at 0.80, with an alpha of 0.05, and a partial eta squared value of 0.055 (effect size f = 0.241). The power analysis indicated that 11 participants per group would be needed. To account for possible participant attrition, the sample size was increased to 14 participants per group (N = 42).

Participants were primarily recruited via flyers posted at approved locations on the University of Wisconsin-Madison campus and at health clinics throughout the greater Madison, WI area. Interested individuals underwent a phone screen to assess potential eligibility. In order to participate in the study, participants from the three groups must have been women between the ages of 18 and 45 years old. This study focused solely on women in order to advance the understanding of the role of exercise in promoting mental health in women with PTSD. Participants from the PTSD group must also have had a current diagnosis of PTSD as assessed via the Clinician-Administered PTSD Scale for DSM-V (CAPS-5; Weathers et al., 2013), which is considered the gold standard in PTSD assessment. Trauma-exposed adults without PTSD must have been exposed to at least one potentially traumatic event that met DSM-V PTSD criteria (assessed via Life Events Checklist for DSM-V, LEC-5; Weathers et al., 2013), but had no lifetime diagnosis of PTSD. Participants from the trauma-free control group (i.e., non-trauma exposed controls, TFC) must not have been exposed to one or more traumatic events based on responses to the LEC-5. Exclusion criteria for participants from all three groups included: being pregnant or planning to become pregnant; having a history of light headedness or fainting during blood draws or physical activity; having a history of chest pain during physical activity; having a bone, joint, or other medical condition that may be worsened by physical activity, having asthma, taking medications for any chronic diseases such as high blood pressure or diabetes, responding ‘yes’ to any of the seven questions on the Physical Activity Readiness Questionnaire; having or receiving treatment for a current or past diagnosis of any psychotic disorder, or any major medical, cognitive, or neurological disorders; or severe major depressive disorder (score > 30 on the Beck Depression Inventory II, BDI-II; Beck, Steer, & Brown, 1996) and/or an indication of suicidality (response of “3” on item 9 of BDI-II).

Study Visits

Participants came into the laboratory for a total of three study visits. The first session was a familiarization session, while study visits 2 and 3 involved experimental procedures. Study visits 2 and 3 occurred at approximately the same time of day for each participant (± 60 min), and visit 3 occurred 48 to 72 hours after study visit 2 (M = 3.23 days; SD = 1.26). Participants were instructed not to eat within two hours nor exercise within 24 hours of testing in order to minimize eCB variations. Participants were compensated ($75) for completing the study (at the end of study visit 3).

Familiarization Session (Study Visit #1).

Written informed consent was obtained at the first session. Participants completed the BDI-II (Beck et al., 1996), and the LEC-5 (Standard Self-Report version; Weathers et al., 2013). Following completion of the BDI-II and LEC-5, participants that indicated they had been exposed to at least one traumatic event were administered the PTSD Checklist for DSM-V (PCL-5; Blevins, Weathers, Davis, Witte, & Domino, 2015). Participants that self-reported meeting criteria for a current diagnosis of PTSD (based on PCL-5 responses; see supplementary material) were administered the CAPS-5. Next, participants completed a neuropsychiatric interview (Mini-International Neuropsychiatric Interview, MINI; Sheehan et al., 1998) in order to classify the sample and ensure participants met inclusion and exclusion criteria. All interview-based assessments (CAPS-5, MINI) were conducted by a trained clinical interviewer. Participants then completed a packet of questionnaires (i.e., Beck Anxiety Inventory, [BAI; Beck & Steer, 1993], Pittsburgh Sleep Quality Index, [PSQI; Buysse, Reynolds, Monk, Berman, & Kupfer, 1989]; Perceived Stress Scale, [PSS; Cohen, Kamarck, & Mermelstein, 1983], and a Menstrual Cycle Log). Next, participants were introduced to the experimental protocols (i.e., the Neutral-Predictable-Unpredictable (NPU) threat task, exercise, and quiet rest) to be administered at study visits 2 and 3. The NPU threat task is a reliable and commonly used psychophysiological task that involves exposure to unpredictable and predictable mild electric shock (rated as very uncomfortable but not painful). Administration of the task allows for the assessment of self-reported anxiety and fear ratings to unpredictable and predictable threats. Therefore, in line with standard and widely used protocols, during the familiarization session, we administered a commonly used shock workup procedure to ensure that participants received a mild electric shock that was rated by each participant as very uncomfortable but not painful (Kaye, Bradford, & Curtin, 2016; Bradford, Magruder, Korhumel, & Curtin, 2014; Schmitz & Grillon, 2012). Specifically, we administered a shock workup procedure that consisted of administering a series of increasing mild electric shocks administered to the fingers (distal phalanges of the second and fourth fingers) of the participant’s non-dominant hand. The shocks had an intensity ranging from 0.5 to 7.0 milliamperes and a duration of 200 ms. Participants were asked to rate the electric shocks to find an upper level that they were comfortable with to be used in the main task. The selected shock was rated by each participant using a 0 (can’t feel shock) to 100 (highest you can tolerate) scale. Participants were not informed that their subjective maximum tolerated shock from this procedures would be used during the NPU threat task during visits 2 and 3. Moreover, participants were provided instructions on the three different conditions, one in which no shock was administered and two during which shocks would be administered either predictably (P) or unpredictably (U). More specifically, participants were instructed that in the P condition, shocks would be administered only in the presence of a cue, and that in the U condition, shocks would be administered at any time (i.e., in the presence or absence of the cue). Lastly, participants were provided with an accelerometer (Actigraph GT3X+, Actigraph, Pensacola, FL) to wear on a belt around their waist for 7 days to determine average levels of physical activity (see supplementary materials for accelerometry data analyses and results). Participants were given instructions on the proper use of the accelerometer and were also asked to complete an activity log in conjunction with wearing the device. Participants returned the accelerometers and activity logs when they returned for their second visit. Further details (e.g., scoring, psychometric properties) on clinical assessments and questionnaires can be found in the supplementary material.

Study Visits 2 & 3.

The second session took place approximately one week after the first study visit. Using a computerized program, participants were randomly assigned to one of two conditions (exercise or quiet rest) for their second session. Participants engaged in the remaining condition at their third session (ideally 48 to 72 hours after study visit 2; see supplementary material for results on visit order effects). Upon arriving at the lab for their second study visit, participants completed questionnaires designed to assess psychological mood states (Profile of Mood States, [POMS; McNair, Lorr, & Droppleman, 1971]; state version of the State-Trait Anxiety Inventory [STAI; Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983]; Positive and Negative Affect Schedule [PANAS; Watson, Clark, & Tellegen, 1988]). Next, participants had their blood drawn (separate sticks) immediately prior to and following completion of either exercise or quiet rest (described below). Immediately after the post-condition blood draw, participants completed the same mood state questionnaires (POMS, STAI, PANAS) and then engaged in the NPU threat task (described below). All participants completed the anxiety and fear rating questionnaire (AFRQ) immediately after completing the NPU threat task at study visits 2 and 3. The AFRQ (which was developed specifically for assessing anxiety and fear outcomes to the NPU task) required participants to retrospectively rate their level of anxiety and fear during the various conditions using 1 (not anxious/fearful) to 5 (extremely anxious/fearful) Likert-type scales (Schmitz & Grillon, 2012; Kaye, Bradford, & Curtin, 2016). Specifically, anxiety ratings were assessed based on how anxious participants felt during the unpredictable shock condition (while the cue was present), and fear ratings were assessed based on how fearful participants felt during the predictable shock condition (while the cue present)(Kaye, Bradford, & Curtin, 2016).

Aerobic Exercise and Quiet Rest.

The acute aerobic exercise session consisted of a 5-minute warm-up of light intensity activity (40–60% age-adjusted maximum heart rate; MHR) on a treadmill followed by walking or running at a moderate intensity (i.e., between 70–75% MHR) for 30 minutes. Participants then cooled down for 5 minutes at a light-intensity. The exercise intensity was selected based on previous research demonstrating a significant increase in eCBs (Crombie et al., 2018; Brellenthin et al., 2017; Raichlen et al., 2012). Ratings of perceived exertion (RPE) and heart rate (HR) were collected every five minutes during exercise. In contrast, the quiet rest session required participants to sit quietly in a sound-dampened chamber for 40 minutes, in order to match for time spent engaging in the aerobic exercise session. Participants wore a HR monitor (Polar, Lake Success, NY) and standardized scripts were used to explain the exercise condition as well as Borg’s RPE scale (6–20; Borg, 1998).

NPU threat task.

During the NPU threat task participants viewed a series of colored square “cues” displayed in the center of a computer screen with a black background. In line with a previously published protocol (Kaye, Bradford, & Curtin, 2016) cues were presented in a blocked design with three conditions: no-shock (N), predictable shock (P), and unpredictable shock (U). Each shock condition was presented twice and separated by no-shock conditions. Condition order was counterbalanced (PNUNUNP, UNPNPNU) both within and between subjects. All blocks included six cues presented for 5s separated by a variable intertrial interval (ITI; ranging from 14 to 20 seconds). Startle probes were administered once during each cue presentation and once during each ITI (no cue period). A white fixation cross remained in the center of the monitor during the cues and ITI. A 200ms electric shock was administered 200 ms prior to the cue offset during every cue in the predictable shock conditions (i.e., appearance of cue predicted shock occurrence). Electric shocks were administered at pseudorandom times during both the cues and ITIs in the U conditions (i.e., occurrence of shock was unpredictable). Shocks occurred 2 or 4.8 seconds post cue onset and 4, 8, or 12 seconds post cue offset in the U condition. A total of six electric shocks were administered in each condition (U, P) and no electric shock was administered during the no-shock condition. Each block lasted approximately 150s, and the entire task lasted approximately 20 minutes. Similar to the familiarization session, participants were provided instructions on the three different conditions (Kaye, Bradford, & Curtin, 2016).

Sample collection, processing, and endocannabinoid assays

Blood draws were carried out while participants were seated. Upon collection, samples (BD Vacutainer, K3E EDTA K3 collection tubes; Greiner Bio‐One, Monroe, North Carolina) were then centrifuged (4 °C at 3500 RPM) within 2 min of collection, separated into aliquots, and frozen at −80 °C until eCB and related lipid extractions took place. Following lipid extraction using a solid phase column, the concentrations of eCBs (AEA and 2-AG), along with related biogenic lipids (PEA, OEA, 2-oleoylglycerol [2-OG]) were quantified using stable isotope-dilution, electrospray ionization liquid chromatography/mass spectrometry (LC-ESI-MS-MS) as previously described (Crombie et al., 2018). PEA and OEA are noncannabinoid fatty-acid ethanolamides that share some biosynthetic and metabolic pathways with AEA and belong to the family of N-acyl ethanolamides, however they are agonists at peroxisome proliferator-activated receptors as opposed to CB1 and CB2 receptors. Similarly, 2-OG is a monoacylglycerol and is a structural analog of 2-AG, as both have overlapping mechanisms of synthesis and degradation (Di Marzo, De Petrocellis, & Bisogno, 2005; Hillard, 2000).

Statistical Analyses

All analyses were conducted with SPSS Version 25.0 for Windows. A series of one-way ANOVAs (for parametric data) and Pearson’s chi-squared tests (for non-parametric data) were conducted (along with post-hoc tests) in order to detect the presence of group differences in baseline characteristics (e.g., age, race, ethnicity, BMI, anxiety, depression, stress), lifestyle factors (e.g., smoking, alcohol, caffeine use), and behavioral outcomes (e.g., physical activity levels, sedentary behaviors). Additionally, independent-samples t-tests were conducted to compare mean differences in trauma-exposure and PTSD symptoms and symptom severity between the trauma-exposed and PTSD groups. Participants from the trauma-free control group did not endorse experiencing any traumatic events and therefore were not included in any trauma and PTSD outcome analyses. Potential group differences in the exercise stimulus variables (i.e., exercising heart-rate, exercising heart-rate percentage of max, % RPE, treadmill speed, and treadmill incline) were examined via a series of one-way ANOVAs. Two-way (group x condition) ANOVAs were conducted to examine mean differences in subjective fear and anxiety ratings to predictable and unpredictable shock during the NPU task following quiet rest and exercise. Circulating concentrations of eCBs and mood states were compared between and within groups using a series of 3 (group: TFC, TE, PTSD) x 2 (condition: QR, EX) x 2 (time: pre, post) mixed-design, repeated measures ANOVAs. The raw values for circulating concentrations of eCBs were non-normal, and therefore concentrations were log-transformed (in order to meet the normality assumption) before conducting further analyses. The overall alpha level was set at = 0.05, respectively, for each set of analyses addressing the three primary hypotheses. Following Fisher’s LSD procedures, if the initial ANOVA yielded a significant interaction, simple effects were calculated for all pairwise comparisons set at α_pc =0.05. Cohen’s d was used to calculate effect sizes in order to provide a quantitative measure of the magnitude (small [d = 0.20 to 0.49]; moderate [d = 0.50 to 0.79]; large [d ≥ 0.80]) of the treatment effect (i.e., exercise, QR) on anxiety and fear ratings, eCBs, and mood state outcomes (Cohen, 1988). Lastly, Spearman’s Rho correlation coefficients were calculated to examine relationships between differences in eCB changes between conditions (i.e., change in eCBs following aerobic exercise – change in eCBs following quiet rest) and differences in fear and anxiety responses to predictable and unpredictable threat between conditions (i.e., fear and anxiety responses during quiet rest – fear and anxiety responses during exercise).

Results

Participant Characteristics

Forty-eight individuals expressed an initial interest in the study and were provided with a detailed explanation of study procedures. Six individuals declined to be taken through the phone screen due to issues/concerns with the blood draws and/or shock procedures. In total, forty-two women enrolled in this study. One individual was excluded after signing informed consent (due to safety concerns) and one individual dropped out after completing the familiarization session. Thus, data was analyzed on a total of forty women (M age = 24.13, SD = 5.95) from three groups: trauma-free controls (n = 12), trauma-exposed women without PTSD (n = 14), and trauma-exposed women with PTSD (n = 14). There were no significant differences (p > .05) between groups for body mass index, race, ethnicity, marital status, education, self-reported overall health, or antidepressant, anxiolytic, or other medication usage. There was a significant group difference in age (F(2,37) = 4.54, p = .017), as the average age for the PTSD group was significantly greater compared to the trauma-free controls (p = .019) but not the trauma-exposed group (p = .083). Additionally, there were significant group differences for baseline anxiety (F(2,37) = 12.36, p < .001), depression (F(2,37) = 12.44, p < .001), perceived stress (F(2,37) = 25.31, p < .001), and sleep quality (F(2,37) = 7.49, p = .002). Post hoc tests revealed that the PTSD group exhibited significantly (ps < .01) greater anxiety, depression, perceived stress, and poorer sleep quality compared to both the trauma-exposed adults and trauma-free controls, with non-significant differences (ps from .536 to .997) between trauma-exposed adults and trauma-free controls (see Table 1).

Table 1.

Baseline participant demographics and characteristics

Variable	Control	Trauma Exposed	PTSD
Age (years)	21.42 ± 3.88	23.00 ± 5.38	27.57 ± 6.60*
Sex (# of women, % of sample)	12 (100%)	14 (100%)	14 (100%)
BMI (kg/m2)	23.04 ± 3.07	22.61 ± 1.77	26.51 ± 7.05
(#, % overweight)	2 (16.6)	1 (7.1)	0 (0)
(#, % obese)	0 (0)	0 (0)	5 (35.7)
Anxiety (BAI)	4.58 ± 7.15	7.57 ± 8.56	22.58 ± 12.39**
Depression (BDI-II)	4.17 ± 4.95	4.36 ± 2.34	19.92 ± 10.88**
Stress (PSS)	9.58 ± 5.99	12.00 ± 4.51	24.92 ± 6.64**
Sleep Quality (PSQI – Global Score)	4.42 ± 2.94	4.93 ± 2.30	9.08 ± 4.4**
Medication
(#, % taking antidepressants)	2 (17)	4 (29)	6 (43)
(#, % taking anxiolytics)	1 (8)	2 (17)	6 (43)
(#, % taking other)	3 (25)	3 (21)	6 (43)
Race
(#, % white)	11 (92)	12 (86)	9 (64)
(#, % Black/African-American)	1 (8)	0 (0)	3 (21)
(#, % Asian)	0	1 (7)	1 (7)
(#, % multiracial)	0 (0)	0 (0)	0 (0)
(#, % other)	0 (0)	1 (7)	1 (7)
Ethnicity
(#, % Hispanic/Latino)	0 (0)	3 (21)	2 (14)
(#, % non-Hispanic/Latino)	12 (100)	11 (79)	12 (86)
Marital Status
(#, % single)	11 (92)	12 (86)	13 (93)
(#, % married)	1 (8)	2 (14)	1 (7)
Education
(#, % high school graduate only)	0 (0)	0 (0)	3 (21)
(#, % some college)	8 (67)	8 (57)	10 (72)
(#, % college graduate)	4 (33)	6 (43)	1 (7)
Overall Health
(#, % excellent)	2 (17)	3 (21)	1 (7)
(#, % very good)	9 (75)	10 (71)	5 (36)
(#, % good)	1 (8)	1 (7)	7 (50)
(#, % fair)	0 (0)	0 (0)	1 (7)

Note. PTSD = Posttraumatic Stress Disorder; BMI = body mass index; BAI = Beck Anxiety Inventory; BDI-II = Beck Depression Inventory-II; PSS = Perceived Stress Scale; PSQI = Pittsburgh Sleep Quality Index. Values listed as M ±SD, unless otherwise noted.

^*significant group difference between PTSD group and healthy control groups, p < .05

^**significant group difference between PTSD group and trauma exposed and healthy control groups, p < .01.

Lifestyle Factors/Behaviors

There were no significant differences (p > .05) between groups for smoking, alcohol, caffeine, or cannabis usage. In terms of activity monitor derived physical activity and sedentary outcomes, there were no significant differences (p > .05) between groups in sedentary behavior time, light-intensity physical activity, moderate-intensity physical activity, vigorous-intensity physical activity, device wear time, and number of days with valid wear time. Data was analyzed on 37 participants, as 3 participants (one from each group) did not meet valid wear-time requirements (see supplementary material and Supplementary Table 1).

Trauma Exposure and PTSD Symptoms and Severity

The mean number of traumatic events (either directly happened to them, witnessed, or combined number of total events) endorsed by the PTSD group was significantly greater (p < .001) than the mean number of traumatic events endorsed by the trauma-exposed group. On average, participants from the PTSD group experienced 13.92 (SD = 7.89) traumatic events over their lifetime. The index events that participants indicated were the most bothersome over the past month (i.e., the events considered while completing the PCL-5 and the clinical interview) all involved interpersonal trauma (physical and/or sexual assault happening to them, n =11; witnessing physical assault, n = 1; experiencing sudden accidental death to a close family member or friend, n = 2). On average, participants from the trauma-exposed group experienced 5.21 (SD = 3.04) traumatic events over their lifetime. The index events for the trauma-exposed group included: life threatening illness to a family member (n = 4), sudden accidental death to a family member/close friend (n = 3), life threatening illness happened to them (n = 2), learning about a close friend being sexually assaulted (n = 2), experiencing a transportation accident (n = 2), and being sexually assaulted (n = 1). In addition to trauma exposure, there were significant differences between groups for total self-reported PTSD symptoms (i.e., number of individual items with severity scores of ≥ 2; t(26) = 9.59, p < .001) and total symptom severity (t(26) = 9.10, p < .001), with a greater number of symptoms and greater symptom severity for the PTSD group compared to the trauma-exposed adults (see Supplementary Table 2). In addition to self-reported outcomes, the PTSD group completed a clinical interview to assess PTSD symptoms and symptom severity (overall and for each cluster; see supplementary Table 2). All participants in the PTSD group met criteria for a current diagnosis of PTSD based on their responses to both the self-reported measure (PCL-5) and clinical interview (CAPS-5).

Exercise Stimulus Variables

There were no significant group differences for average percent of HR max (F(2,36) = 2.12, p = .134), RPE (F(2,36) = .335, p = .717), and treadmill incline (F(2,36) = .91, p =.411) throughout the exercise session, with significant group differences for average exercising HR (F(2,36) = 6.52, p = .003) and average treadmill speed (F(2,36) = 7.68, p = .002). On average, participants exercised at a moderate-intensity as assessed via percent of HR max and via RPE (see Table 2). The PTSD group had a lower exercising HR compared to the trauma-exposed group (p < .01) but not the trauma-free control group (p > .05). There was no significant difference in exercising HR between trauma-exposed adults and trauma-free controls (p > .05). Additionally, the PTSD group exercised at a slower treadmill speed compared to the trauma exposed group (p < .001) and the trauma-free controls (p < .05). There was no significant difference in treadmill speed between trauma-exposed adults and trauma-free controls (p > .05).

Table 2.

Means and standard deviations for exercise session variables

Variable	Control	Trauma Exposed	PTSD	Total
Ratings of perceived exertion (RPE)	12.01 ± 1.30	11.57 ± 1.35	11.82 ± 1.50	11.79 ± 1.36
Exercising heart rate (bpm)	146.62 ± 5.51	148.61 ± 5.01	140.44 ± 8.85*	145.28 ± 7.40
Heart rate percentage (% MHR)	73.82 ± 2.01	75.10 ± 2.51	73.08 ± 3.06	74.03 ± 2.65
Treadmill speed (km/h)	6.82 ± 0.60	7.41 ± 1.31	5.63 ± 1.43**	6.63 ± 1.38
Treadmill incline (% grade)	1.49 ± 1.33	0.96 ± 0.74	1.35 ± 1.04	1.26 ± 1.05

Note. PTSD = Posttraumatic Stress Disorder; bpm = beats per minute; MHR = maximum heart rate; km/h = kilometers per hour.

^*significant difference from trauma exposed group, p < .01

^**significant difference from trauma exposed and healthy control groups, p < .01.

Anxiety and Fear Ratings to NPU Threat Task Following Exercise and Quiet Rest

There were significant condition main effects for anxiety (F(1,37) = 14.856, p < .001) and fear (F(1,37) = 8.214, p = .007) ratings to unpredictable and predictable threat, respectively, indicating that anxiety and fear ratings were significantly lower for the overall sample following exercise compared to quiet rest (see Figures 1–2; Table 3). There were non-significant (p > .05) group and group x condition interactions for anxiety and fear ratings to unpredictable and predictable threat.

Figure 1.

Anxiety ratings to NPU threat task following quiet rest and exercise for trauma-free controls, trauma exposed, and PTSD groups, and overall sample. Ratings in response to how anxious participants felt during the unpredictable shock condition (cue present). Higher ratings are indicative of greater anxiety. Anxiety ratings were significantly lower following exercise compared to quiet rest (denoted*).

Figure 2.

Fear ratings to NPU threat task following quiet rest and exercise for trauma-free controls, trauma exposed, and PTSD groups, and overall sample. Ratings in response to how fearful participants felt during the predictable shock condition (cue present). Higher ratings are indicative of greater fear. Fear ratings were significantly lower following exercise compared to quiet rest (denoted*).

Table 3.

Means, standard deviations, and effect sizes for anxiety and fear ratings to NPU threat task following quiet rest and exercise

	Quiet Rest	Exercise
Outcome	M ± SD	M ± SD	Cohen’s d
Anxiety ratings
Control	4.00 ± 0.85	3.33 ± 0.98	0.73
Trauma Exposed	3.29 ± 1.20	3.00 ± 1.24	0.24
PTSD	4.15 ± 0.99	3.46 ± 0.88	0.74
Overall	3.79 ± 1.08	3.26 ± 1.04	0.50
Fear ratings
Control	3.50 ± 0.67	3.08 ± 1.00	0.49
Trauma Exposed	3.21 ± 1.12	2.93 ± 1.38	0.23
PTSD	3.77 ± 1.24	3.15 ± 1.07	0.54
Overall	3.49 ± 1.05	3.05 ± 1.15	0.40

Note. Ratings were obtained from AFRQ questionnaire administered immediately following the NPU threat task at both visits. Responses were rated on Likert scale ranging from 1 (not at all anxious/fearful) to 5 (very anxious/fearful). Anxiety ratings were based on responses to the unpredictable shock condition (cue present), and fear ratings were based on responses to the predictable shock condition (cue present).

Circulating Endocannabinoid Responses

There were significant (p < .05) condition x time interactions for AEA (F(1,37) = 43.017, p < . 001), 2-AG (F(1,37) = 9.925, p = . 004), OEA (F(1,37) = 11.932, p = .002), and PEA (F(1,37) = 5.299, p = . 029). Circulating concentrations of AEA, 2-AG, OEA, and PEA increased significantly from pre- to post-exercise (ps = <.001, .011, <.001, .001, respectively), but not following quiet-rest (ps = .114, .194, .215, .237 respectively) for the entire sample (see Figure 3; Supplementary Table 3). All other main effects and interactions were non-significant (p > .05).

Figure 3.

Means and standard errors for circulating concentrations of AEA [A] and 2-AG [B] before and after quiet rest and exercise for all three groups (trauma-free control, trauma exposed, and PTSD) and for overall sample. PTSD = Posttraumatic stress disorder; AEA = N-arachidonoylethanolamine; 2-AG = 2-arachidonoylglycerol. *significant condition x time effect, p < .01.

Mood State Responses

There were significant condition x time interactions for fatigue (F(1,37) = 4.983, p = . 032), confusion (F(1,37) = 4.116, p = . 049), total mood disturbance (F(1,37) = 7.265, p = . 011) and positive affect (F(1,37) = 16.540, p < . 001) for the overall sample. Analysis of simple effects indicated significant reductions in fatigue (F(1,37) = 23.248, p < . 001), confusion (F(1,37) = 9.907, p = . 003), and total mood disturbance (F(1,37) = 39.136, p < . 001), and significant increases in positive affect (F(1,37) = 10.109, p < . 001) for the overall sample following exercise but not quiet rest (ps = .229 to .526). There were significant group x time interactions for tension (F(2,37) = 3.566, p = . 038), anger (F(2,37) = 4.539, p =. 039), negative affect (F(2,37) = 3.959, p < . 05), and state anxiety (F(2,37) = 4.688, p < . 05), with analysis of simple effects indicating that the PTSD group experienced significant reductions in tension (F(1,37) = 25.688, p < . 001), anger (F(1,37) = 18.789, p < . 001), negative affect (F(1,37) = 17.306, p < . 001), and state anxiety (F(1,37) = 17.766, p < . 001) following exercise and quiet rest. Finally, there were significant group x condition x time interactions for vigor (F(2,37) = 3.752, p =. 033) and depression (F(2,37) = 7.372, p = . 002). Analysis of simple effects indicated significant decreases in vigor following quiet rest for the trauma-free control (F(1,37) = 7.220, p = . 010) and trauma exposed groups (F(1,37) = 4.649, p = . 037), with significant increases following exercise for the trauma-free control (F(1,37) = 16.133, p < . 001) and PTSD (F(1,37) = 5.997, p = . 019) groups. Depression scores were significantly reduced (F(1,37) = 65.472, p < . 001) following exercise for the PTSD group. See Table 4 for means, standard deviations, and effect sizes.

Table 4.

Means and standard deviations for state anxiety and mood state responses prior to and following quiet rest and exercise

	Quiet Rest Session		Exercise Session
	Pre	Post	Pre	Post
Mood Outcomes	M ± SD	M ± SD	M ± SD	M ± SD
State Anxiety
Control	30.25 ± 7.16	28.25 ± 5.07	31.25 ± 9.13	28.67 ± 6.29
Trauma Exposed	33.79 ± 13.07	33.64 ± 12.76	33.00 ± 7.51	33.14 ± 8.41
PTSD	45.54 ± 8.84	41.85 ± 11.10	46.23 ± 11.45	37.23 ± 9.24
Overall	36.62 ± 11.87	34.72 ± 11.53	36.87 ± 11.42	33.13 ± 8.64
Tension
Control	4.50 ± 2.78	2.42 ± 2.64	4.00 ± 3.79	3.42 ± 3.06
Trauma Exposed	7.43 ± 6.93	5.43 ± 8.24	4.79 ± 3.33	3.86 ± 3.61
PTSD	11.23 ± 7.00	8.23 ± 5.60	11.54 ± 5.58	6.54 ± 3.15
Overall	7.79 ± 6.45	5.44 ± 6.38	6.79 ± 5.42	4.62 ± 3.50
Depression
Control	0.83 ± 1.11	0.67 ± 1.23	1.58 ± 2.50	0.67 ± 1.30
Trauma Exposed	2.86 ± 7.05	1.93 ± 6.65	1.71 ± 2.73	1.36 ± 2.27
PTSD	10.69 ± 11.36	10.46 ± 12.41	14.30 ± 13.43	7.15 ± 10.79
Overall	4.85 ± 8.74	4.38 ± 9.13	5.87 ± 9.89	3.08 ± 6.90
Anger
Control	0.50 ± 1.00	0.08 ± 0.29	0.50 ± 1.00	0.33 ± 0.78
Trauma Exposed	1.36 ± 1.91	0.79 ± 1.53	1.14 ± 1.70	1.00 ± 1.96
PTSD	4.31 ± 5.27	2.77 ± 2.55	5.23 ± 5.34	2.38 ± 3.20
Overall	2.08 ± 3.60	1.23 ± 2.05	2.31 ± 3.84	1.26 ± 2.34
Confusion
Control	3.33 ± 2.23	2.75 ± 1.48	3.08 ± 1.56	2.83 ± 2.44
Trauma Exposed	4.00 ± 4.00	4.14 ± 3.57	3.86 ± 2.35	3.21 ± 2.33
PTSD	8.77 ± 4.66	8.38 ± 3.62	9.38 ± 5.52	6.46 ± 3.95
Overall	5.38 ± 4.44	5.13 ± 5.77	5.46 ± 4.49	4.18 ± 3.35
Fatigue
Control	3.92 ± 3.06	2.92 ± 2.71	3.50 ± 2.84	1.75 ± 2.18
Trauma Exposed	4.21 ± 2.83	4.21 ± 3.42	4.14 ± 4.24	2.79 ± 2.61
PTSD	10.08 ± 6.51	8.92 ± 8.06	9.31 ± 5.41	4.77 ± 3.47
Overall	6.08 ± 5.20	5.38 ± 5.77	5.67 ± 4.96	3.13 ± 3.01
Vigor
Control	15.75 ± 5.80	13.25 ± 6.27	14.83 ± 6.74	19.17 ± 6.45
Trauma Exposed	13.21 ± 3.93	11.36 ± 4.94	14.00 ± 6.75	14.07 ± 7.36
PTSD	7.85 ± 4.40	7.31 ± 4.23	10.85 ± 7.09	13.38 ± 7.56
Overall	12.21 ± 5.65	10.59 ± 5.61	13.21 ± 6.90	15.41 ± 7.43
TMD
Control	97.33 ± 11.08	95.58 ± 10.96	97.83 ± 13.35	89.83 ± 13.98
Trauma Exposed	106.64 ± 21.19	105.14 ± 22.22	101.64 ± 15.34	98.14 ± 14.01
PTSD	137.23 ± 30.54	131.46 ± 30.60	138.92 ± 31.40	113.92 ± 23.12
Overall	113.97 ± 27.86	110.97 ± 27.02	112.90 ± 28.17	100.85 ± 19.80
Negative Affect
Control	10.67 ± 1.56	10.42 ± 0.79	11.00 ± 1.86	10.33 ± 0.89
Trauma Exposed	13.57 ± 6.78	13.29 ± 5.94	12.36 ± 2.59	11.79 ± 2.33
PTSD	16.38 ± 5.63	14.31 ± 5.25	16.46 ± 9.10	13.38 ± 4.31
Overall	13.62 ± 5.64	12.74 ± 4.86	13.31 ± 5.91	11.87 ± 3.08
Positive Affect
Control	30.50 ± 8.19	28.08 ± 6.52	30.75 ± 10.17	34.25 ± 8.99
Trauma Exposed	28.36 ± 6.89	25.71 ± 6.38	28.79 ± 8.65	28.50 ± 9.53
PTSD	22.31 ± 8.46	21.85 ± 9.50	23.38 ± 9.62	27.00 ± 10.60
Overall	27.00 ± 8.39	25.15 ± 7.83	27.59 ± 9.72	29.77 ± 9.98

Note. Values listed as M ± SD. PTSD = Posttraumatic Stress Disorder; TMD = total mood disturbance. Tension, depression, anger, confusion, fatigue, vigor, and total mood disturbance scores were obtained from the Profile of Mood States (POMS) questionnaire. Positive and Negative affect scores were obtained from the Positive and Negative Affect Schedule (PANAS). State Anxiety scores were obtained from the state anxiety subset of 20 items from the State-Trait Anxiety Inventory (STAI). TMD was derived by adding 100 to the sum of the negative mood state subscales (i.e., depression, tension, confusion, fatigue, anger) minus the positive mood state subscale (i.e., vigor).

Associations between endocannabinoids and anxiety and fear ratings

There were significant positive correlations between the difference in 2-AG changes between conditions and the difference in anxiety (ρ = .408, p = .023) and fear (ρ = .388, p = .031) responses between conditions, suggesting that those with greater increases in 2-AG had greater differences in anxiety and fear ratings between conditions (i.e., greater reductions following exercise compared to quiet rest). There were no significant relationships between differences in AEA changes between conditions and differences in anxiety (ρ = .219 p = .236) or fear (ρ = .201 p = .278) ratings between conditions.

Discussion

The current study found that anxiety and fear ratings to unpredictable and predictable threat were significantly lower following aerobic exercise compared to quiet rest for the total sample of women which included women without a trauma history, as well as trauma exposed women with and without a current diagnosis of PTSD. To date, there has only been one prior investigation examining anxiety and fear responses to unpredictable and predictable threatening stimuli following aerobic exercise (Lago et al., 2018). Lago and colleagues (2018) administered the NPU task and found that an acute bout of aerobic exercise (30 minutes of cycling at 60–70% heart rate reserve) in 34 healthy male and female young adults (M age = 26 yrs), significantly reduced startle potentiation during the unpredictable condition (i.e., anxiety-potentiated startle), but not during the predictable condition (i.e., fear-potentiated startle), compared to a control condition. Although Lago et al. reported small effect size changes (between conditions) in self-reported anxiety and fear ratings to the task, it is important to note that participants from that study also completed an additional task (following exercise) involving shock administration prior to completing the NPU task, thus making it difficult to determine if the additional task had an influence on self-reported ratings (Lago et al., 2018). Furthermore, participants in the current study provided retrospective anxiety and fear ratings immediately after the task, whereas participants from Lago et al. provided ratings continuously throughout the task. Although the aforementioned study was the most similar in terms of examining anxiety and fear responses to a threat task following exercise and quiet rest, there have been a few other investigations in adult men and women with and without anxiety disorders providing evidence in favor of acute aerobic exercise to serve as a protective stress-buffer to a variety of lab-based stressors (i.e., cholecystokinin-4 (CCK-4) administration, inhalation of 35% CO₂) when the stressor is administered shortly after the exercise bout has ended (Esquivel, Schruers, Kuipers, & Griez, 2002; Esquivel, Diaz-Galvis, Schruers, & Berlanga, 2008; Rejeski, Thompson, Brubaker, & Miller, 1992; Smits, Meuret, Zvolensky, Rosenfield, & Siedel, 2009; Ströhle, Feller, Strasburger, Heinz, & Dimeo, 2006; Ströhle et al., 2009).

The current study also examined the eCB system as a potential mechanism underlying the anxiolytic effects of exercise. Circulating concentrations of eCBs (AEA and 2-AG) increased significantly in response to moderate-intensity aerobic exercise but not following quiet rest. Although all three groups experienced a significant increase, both the trauma-free control and trauma exposed groups exhibited large effect size increases in AEA following exercise compared to a moderate increase in the PTSD group. The current study also found that individuals with greater differences in 2-AG between conditions (exercise and quiet rest) reported greater differences in both anxiety and fear responses between conditions. In other words, those with a greater exercise-induced change in circulating 2-AG tended to exhibit a greater exercise-induced reduction in fear and anxiety ratings to the NPU task. Although the eCB system’s modulatory role in reducing anxiety and fear responses to stressors is not fully understood, these findings are in line with a large amount of literature suggesting that increased circulating concentrations of eCBs in response to stress, appear to play an important and protective, stress-buffering role (e.g., prevent motion sickness, nausea, vomiting, reduce subjective anxiety and stress ratings to stressors; greater tolerance to stressor; Choukèr et al., 2010; Schroeder et al., 2009; Stricht, Rock, Limebeer, & Parker, 2015). In contrast to exercise, very limited research has been conducted examining circulating eCB responses following quiet rest. In fact, to our knowledge there has only been one other study, which similarly found no significant changes in AEA or 2-AG following three 30-minute sessions of quiet rest in treatment seeking adults with a current substance use disorder diagnosis (Brellenthin, Crombie, Hillard, Brown, & Koltyn, 2019). Dlugos and colleagues (2012) also reported that eCBs did not significantly increase in healthy men and women following a slightly different control condition (i.e., speaking with research assistant about neutral topics including interest and hobbies) compared to a stress condition (i.e., administration of Trier Social Stress Test; Dlugos, Childs, Stuhr, Hillard, & de Wit, 2012).

In addition to influencing anxiety and fear ratings to unpredictable and predictable threats, the current study found that an acute bout of moderate-intensity aerobic exercise was able to elicit significant mood state improvements for women, particularly women with PTSD stemming from interpersonal trauma. Specifically, there were large effect size improvements in tension, fatigue, state anxiety, and total mood disturbance; moderate improvements for depression, anger, and confusion; and small improvements for vigor, negative affect, and positive affect. The reported mood state improvements (i.e., reductions in negative mood states and increases in positive mood states) for the PTSD group are in line with previous investigations examining acute mood state responses to aerobic exercise in clinical populations including, but not limited to: PTSD (Crombie et al., 2018; Crombie et al., 2019), depression (Meyer et al., 2019; Bartholomew, Morrison, & Ciccolo, 2005), panic disorder (Herring et al., 2019), and substance use disorders (Brellenthin et al., 2019). Consistent with previous research in non-clinical populations (Herring et al., 2019; Crombie et al., 2018; Brellenthin et al., 2017; McDowell, Campbell, & Herring, 2016), effect size calculations revealed that the trauma-free control group experienced small to moderate improvements for all mood state outcomes except for confusion, while the trauma exposed group experienced small improvements in tension, confusion, fatigue, vigor, total mood disturbance, and negative affect. Additional research is needed to examine not only acute psychological responses to exercise, but also mental health outcomes following chronic exercise training. Although initial investigations have only recently begun, results are promising as psychological improvements (i.e., reduced depression, anxiety, stress) and reduced PTSD symptomology (number of symptoms and symptom severity) were reported following a 12-week resistance exercise training program in adult men and women with PTSD (Rosenbaum et al., 2015). Moreover, reduced PTSD symptoms were reported in a sample of older adult veterans with PTSD following a multi-component (aerobic, strength, balance, and flexibility) exercise intervention (Hall et al., 2020). Finally, avoidance and hyperarousal symptoms were reduced, and sleep quality improved following a 3-week resistance training program in non-treatment seeking adult women who screened positive for PTSD (Whitworth et al., 2019).

Following quiet rest, the PTSD group experienced significant reductions in tension, anger, total mood disturbance, negative affect, and state anxiety. In contrast, the trauma-free control and trauma exposed groups experienced significant reductions in positive affect and vigor. Although the PTSD group experienced a greater number of mood state improvements following quiet rest compared to the other two groups, effect size calculations indicated that mood state improvements were of a larger magnitude following exercise. Similarly, for the overall sample, exercise led to a greater number of mood state improvements in comparison to quiet rest. Previous research directly comparing acute bouts of aerobic exercise and quiet rest have often found that both conditions result in mood state improvements, although the mood state improvements typically last longer following exercise compared to quiet rest (Brellenthin et al., 2019; Bartholomew et al., 2005; Raglin & Morgan, 1987).

The current study is not without limitations. One of the main limitations is that we relied on self-reported fear and anxiety ratings to the NPU task, as opposed to arguably more objective responses (i.e., startle responses recorded via electromyography). While there is some debate regarding the degree in which behavioral and physiological fear and anxiety responses are correlated and represent an integrated response (Fanselow & Pennington, 2017; Thyer, Papsdorf, Davis, & Vallecorsa, 1984; LeDoux & Pine, 2016) subjective experiences alone provide valuable information that should not be neglected in absence of physiological data (Fanselow & Pennington, 2017). Additionally, other than index events, each individual traumatic event endorsed (via LEC-5) by participants was not verified for meeting Criterion A, which should be noted when interpreting the number of traumatic events experienced by participants. Another limitation is that medication usage was only documented (and not excluded for), and therefore anxiolytic medication usage could have influenced the psychobiological outcomes. However, anxiolytic medication usage did not differ between groups, and individuals that were taking anxiolytic medication exhibited similar mood, anxiety, and fear responses compared to those who were not taking anxiolytic medication. Additionally, we examined the eCB system solely via circulating concentrations. Although we did not measure central eCB system responses, there is preclinical evidence to suggest that increases in peripheral concentrations can influence central process (e.g., brain reward processes; Justinova et al., 2005; Justinova et al., 2011). Another limitation is that fluid intake (during the exercise and quiet rest session) and plasma volume was not measured in this study, and therefore changes in hydration status and plasma volume may have influenced changes in circulating eCB concentrations (Kargotich et al., 1997). Also, although we did not control for and schedule participants based on their menstrual cycle phase, 92% of the analyzed sample had both visits occur in the same phase of the menstrual cycle which mitigates, but does not eliminate, this concern. Future exercise studies examining fear responses (especially studies examining fear extinction) in women should control for menstrual cycle phase (Cover, Maeng, Lebron-Milad, & Milad, 2014; Hammoud, Foa, & Milad, 2020; Sartin-Tarm et al., 2020). Relatedly, it is not clear how these results would generalize to men, as the current study only tested women.

Conclusions

Collectively, the results obtained from this study suggest that exercise is a safe, effective, and beneficial stimulus for improving numerous psychological outcomes in women with PTSD. Specifically, our findings suggest that: 1) anxiety and fear ratings to unpredictable and predictable threats (encountered shortly after bouts of exercise and quiet rest) are lower following moderate-intensity aerobic exercise in comparison to quiet rest; 2) mood states are improved following moderate-intensity aerobic exercise in women including women with PTSD; 3) eCBs increase in response to moderate-intensity aerobic exercise, and 4) exercise-induced increases in circulating eCBs appear to be associated with changes in anxiety and fear ratings to threats. Future research should continue examining mental health outcomes (and potential mechanisms responsible for psychological improvements) in women with PTSD following both acute and chronic exercise.

Highlights:

• Aerobic exercise reduced anxiety and fear ratings to unpredictable and predictable threats.

• Circulating concentrations of endocannabinoids increased following aerobic exercise.

• Mood states improved following aerobic exercise in women with PTSD.

• Aerobic exercise exerts psychological benefits in women with PTSD.