SPECIAL ISSUE: Male-Female Differences in Psychophysiological Stress Profiles Before and After a Group Relaxation/Biofeedback Stress Management Program

Clients

Participants were representative of a university-based community clinic where most people were not under medical supervision and are more fully described in Wilson et al. (2023). They paid a minimal fee in return for their cooperation in completing personality inventories, attending at least 80% of the relaxation training sessions, participating in two psychophysiological stress profiles, and tracking symptoms, medication use, and home practice. In this sample of seven consecutive group classes, there were 141 clients who completed pre-post psychological inventories and stress profiles in addition to attending the group stress management program. There was no significant difference in age between the 80 women (mean = 32.3 years) and 61 men (mean = 29.6 years), who ranged in age from 15 to 67 years. The clients were university students (57%) and community members (43%), and they reported cognitive (36%) and/or somatic (54%) stress symptoms or wished to learn self-regulation skills (10%).

Program

The community-based program consisted of an introductory meeting, then 10 one-hour training sessions across 5 weeks with a follow-up session after 1 month, and was previously described by Wilson et al. (2023). Individual temperature biofeedback was included in the last 5 classes using small, inexpensive portable digital thermometers. Daily home practice was highly encouraged each week, with logs submitted every week. A detailed program outline is at https://www.selfregulationskills.ca/programs/12-session-outline/.

Protocol

During the week preceding the relaxation training program, a 30-minute psychophysiological stress profile was administered in a sound-attenuated laboratory at a room temperature of 72°F. An Autogen 1700 electromyograph collected data on surface muscle tension (sEMG) using a narrow 100- to 200-Hz bandpass, and an Autogen 2000b collected skin temperature, whereas heart rate was assessed with a finger clip photoplethysmograph placed on the nondominant index finger. The sEMG sensors were placed on the middle one-third of the belly of the right upper trapezius muscle, and the ground electrode was placed on the spine to measure muscle tension levels. The trapezius muscle was chosen because it is responsive to mental and physical stressors (Krantz et al., 2004). Fingertip temperature was measured with a thermistor placed under mesh screening that was taped on the middle finger of the nondominant hand. The data were collected, integrated, and stored in a specialized software program (Hoare et al., 1986).

This physiological stress profile procedure was repeated in the week following the 10-session training program. No stress profile data were shared with the participants.

Unlike most stress profiles, these clients were lying supine throughout the profile. The profile events were: EO = baseline – eyes open (5 minutes), EC = baseline − eyes closed (5 minutes), ANT = anticipatory stress after announcement of an upcoming test – eyes closed (2 minutes), COG = cognitive stressor (anagram solving) − eyes open (2 minutes), REC = recovery − eyes closed (5 minutes), EMOT = emotional stressor imagining a personal stressor − eyes closed (2 minutes), REC = recovery – eyes closed (5 minutes), PHY = physical stressor (sudden, loud noise) − eyes closed (2 minutes), and REC = recovery − eyes closed (3 minutes).

Skin Temperature

Skin temperature values were negatively skewed, but applying a power 9 transformation resulted in an approximately normal data distribution, thus meeting statistical assumptions for the use of ANOVA.

Skin temperatures increased significantly from pre- to postprogram for both females (M = 1.94, p = .028) and males (M = 1.64, p = .05) (see Figure 1). Males and females were significantly different from each other in their temperatures, F(1, 128) = 1.29, p = .01, but males and females were not significantly different from each other in the amounts their temperature changed.

View Figure

• Download as Powerpoint
• Download Figure

Stress responses and recovery are very often different and of differential sensitivity to various stressors (Arena & Schwartz, 2016); thus, we further investigated possible differences in responses of males and females to the stressors using a Mann-Whitney U test for both the reaction and then the recovery for each of the stressors. A Bonferroni correction was applied to adjust for the repeated measures to determine the appropriate level of significance.

Females showed significantly larger decreases in temperature than males did during the physical stressor (p = .002) and less recovery from that physical stressor (p = .022) in the preprogram profile, and they showed significantly larger decreases than males during the emotional stressor (p = .015) and physical stressor (p = .010) in the postprogram profile. Females also displayed less temperature recovery following the emotional (p = .015) and physical (p = .001) stressors in the postprogram profile. Males had a more significant decrease than females did during the postprogram cognitive stressor (p = .021). In essence, females’ temperature decreased significantly more than that of males to emotional and physical stressors in the pre- and postprofiles. Males’ temperature decreased significantly more than that of the females only during the cognitive stressor postprogram.

Heart Rate

There were no significant heart rate differences between pre- and postintervention for females or males. Females had significantly higher heart rates than males did, both pre- and postprogram, F(1, 110) = 13.39, p < .001, but males and females did not differ significantly in the amounts their heart rate changed (see Figure 2). Although not statistically significant, there was a trend toward females displaying dampened heart rate response to all stressors in the postprogram profile.

View Figure

• Download as Powerpoint
• Download Figure

sEMG

Both males and females had very low upper trapezius sEMG levels, which may be partially attributed to narrow band sensors (Shaffer & Neblett, 2010) but more likely due to participants lying down during the stress profiles. Removing five outliers resulted in the sEMG values showing a normal distribution.

Although there was a pattern of females showing higher sEMG values than males, it was not significantly different across the entire profile, before and after the program (see Figure 3). Neither males nor females were significantly different in the amount they changed from pre- to postprogram.

View Figure

• Download as Powerpoint
• Download Figure

Further analyses of the patterns of sEMG responses to the stressors and their recoveries were assessed using the Mann-Whitney U test with a Bonferroni correction for significance level. There was a trend toward females displaying more EMG reactivity during all stressors, with statistically significantly more muscle reaction for females in the preprogram emotional stressor (p = .009) and the postprogram cognitive stressor (p = .020). There were no significant male-female differences during the recovery periods.

Psychophysiological Interrelationships

A Rho correlation coefficient assessed the relationship between the physiological and psychological data of subjects for both pre- and postintervention profiles. Trait anxiety (STAI) was significantly positively correlated with females’ preprogram (rho = .32, p < .01) and postprogram (rho = .25, p < .05) sEMG levels, indicating that females with higher trait anxiety displayed higher muscle tension.

More Versus Less Successful

Following the program, those who rated their improvement in well-being as 100% or greater (36% of the sample) were considered the most successful group, whereas those who reported feeling less than 50% improvement (20% of the sample) were considered the least successful group. Statistical analysis on success by gender was not conducted due to the small sample size, but the pattern of responses was reviewed. In both the pre- and postintervention profiles, the most successful females had a pattern of warmer hands and lower sEMG, but a higher heart rate, compared with the least successful females. The most successful males had a pattern of higher sEMG and heart rate and equal temperature in the preprogram profile compared with the least successful males. In the postprogram profile, both male groups had a pattern of equal sEMG levels, but the least successful males had a lower heart rate and warmer hands. These preliminary data suggest that when analyzing the degree of success of a stress management program, males and females appear to have different physiological responses.

Pre-Post Changes in Physiological Measures

The two most significant findings from this study are the decrease in anxiety and the increase in fingertip temperature following a biofeedback-assisted relaxation and stress management program. This is unlike Mauss et al. (2004) and Chang (2011), who found anxiety decreases were accompanied by no physiological changes immediately following relaxation programs.

Skin Temperature

Arena and Hobbs’ (1995) analysis of stress profiles found that with the exception of frontal EMG during a cold pressor stressor, stress profiles are stable over time. Because the profiles were collected throughout the year, the postprogram increase in temperature is not likely due to time-of-year effects but more likely due to the training effect from the program, especially because temperature biofeedback was practiced.

Increased temperature in the postprogram profile for both males and females may be due to the program techniques (breathing, autogenic, progressive muscle relaxation, and alphagenics may have a larger effect on temperature) and that temperature biofeedback training was done within the program. This biofeedback, alone or in combination with other program components, could have provided more skill acquisition or motivation and a physiological effect limited specifically to finger temperature (Wilson et al., 2004; Peper et al., 2015).

Males having warmer hands and lower heart rates than females, which might simply be due to biological differences. The finding that the fingertip temperature of males was consistently warmer than that of females corresponds with previous research (Lucas & McIlvaine, 1985; Sargent et al., 1986; Montgomery, 1988). Blanchard et al. (1989) cited their group mean for 56 combined normal males/females as 88.6°F during the stress profile baseline (eyes closed), which is similar to the 88.2°F mean for the 74 females in this sample; however, the 60 males (eyes closed) had a mean temperature of 91.6°F at baseline. Blanchard et al. (1989) found that individuals with vascular headaches had significantly lower hand temperatures across stress profile conditions, and in this sample, twice as many women as men had vascular headaches. Interestingly, Roberts and McGrady (1996) found 17 white males in medical school had cooler hands (79.8°F) than 10 white females (84.6°F) prior to biofeedback training, but these temperatures were not different after training. Generally, females have colder hands due to hormonal influences, differences in vascularization, and perhaps other factors (Violani & Lombardo, 2003). Because females switch from having warmer to cooler shoulder skin temperatures than males do at about age 16–17 years (Pronina et al., 2015), we wonder if others have found a switch in female hand temperature before and after puberty. Documenting the sex and age of subjects may be important for database accuracy.

The pattern of temperature responses during the profiles is interesting. Females showed significantly larger temperature decreases during the physical stressor than males did in the preprogram profile and during the emotional and physical stressors in the postprogram profile. Skin temperature decreased most in males during the cognitive stressor (and it was significantly different from the females’ response in the postprogram stress profile). Our clinical experience matches these data in that males tend to respond more strongly to the cognitive stressor, be it anagram, Stroop test, math, and so forth, than do females.

Recovery of temperature following emotional and physical stressors in the postprogram profile was less for the females compared with the males. Temperature change has long been associated with emotion (from Ziegler & Cash, 1938, to De Zorzi et al., 2021), and it is also associated with concentration (Serrano-Mamolar et al., 2021). Willmann et al. (2012) found that male finger temperature recovery following a moderate cognitive stressor seems to be related to trait anxiety level. Brosschot et al. (2006) noted that perseverative cognition (repeated or chronic activation of core cognitive processes involved in worry and ruminating) “expands the temporal duration of a stressor beyond the traditional reactivity period to include anticipation and recovery, thereby being the source of prolonged physiological activation and somatic complaints” (p. 116).

Perhaps the differential patterns reflect gender differences in the amount of anxiety, stress, emotion, concentration, perseverative cognition, or performance during the stressor. Evans and Steptoe (2003) stated that “physiological and affective stress reactions in males and females are determined in part by traditionally gender-related psychological characteristics, with greater reaction when the situational demands are not congruent with preferred modes of behavior” (p. 756). It is important for future work to look separately at male and female differences with respect to such variables and how they affect anticipation, stressor response, and recovery from stressors. When selecting and interpreting stress profiles, the stressors need to be clearly identified as there are differences in responses between males and females and perhaps other characteristics related to culture, age, and experience.

The stress management program significantly decreased anxiety, increased self-reported well-being, and showed increased fingertip temperature for both males and females. Pierpaolo et al.’s (2022) training in biofeedback and relaxation for high-level male managers showed improvements not only in psychophysiology (temperature, skin conductance) but also in decision making under stress, suggesting biofeedback-assisted relaxation training may have real-world impacts in stressful decision-making situations.

Heart Rate

Because males show a significantly lower heart rate than females do in laboratory studies (von Scheele et al., 2005; Koenig & Thayer, 2016; Rattel et al., 2020) and home environments (van Kraaij et al., 2020), our similar findings were expected. Although some of the difference may be due to genetics and fitness levels, Dua and King (1987) found a trend of consistently higher heart rates in self-reported worriers across all stress profile conditions. Although worry was not measured in this study, it perhaps contributed to the higher female heart rates.

Although McAdoo et al. (1990), Baum and Grunberg (1991), and Ordaz and Luna (2012) reported that females display more heart rate reactivity during all stressors, that was not found here. Perhaps this is because of the complexity of stress reactivity and recovery. Interacting physiological and psychological processes, as well as factors such as fitness (Forcier et al., 2006; Jackson & Dishman, 2006); reciprocal coupling of the sympathetic and parasympathetic nervous system (Weissman & Mendes, 2021); optimism, autonomy, and mastery (DuPont et al., 2020); worry (Castaneda & Segerstrom, 2004); rumination (Watkins & Roberts, 2020); rehearsing failures, and engaging in self-denigrating thoughts (Roger & Jamieson, 1988) were not accounted for in this study.

Although not statistically significant, there was a trend toward females displaying dampened heart rate response to all stressors in the postprogram profile. Future research should look at whether this heart rate dampening postprogram is a real effect from training. If it holds true, it is then important for self-regulation because heart rate is easily trained with breathing. It may be easier to teach females because of their higher heart rate initial values and responsiveness. Respiration measures should be taken and looked at in relationship with heart rate, which may be easier now due to wearable devices and sensors.

sEMG

Females displayed a pattern of higher EMG (pre- and postprogram) compared with males, similar to the findings of Krantz et al. (2004), who reported that females had higher mean sEMG during baseline and recovery periods. The lack of statistical significance in this sample is possibly because of the small surface area measured, initial low sEMG levels, high degrees of variability, or supine position. It is possible that the females had a tendency to breathe more thoracically than the males did, although this was not investigated. The failure of obtaining significant postprogram changes in EMG and heart rate may be due to the relaxation techniques not affecting them as much, not enough time to have an impact, or the absence of heart rate or EMG biofeedback.

There was a pattern of females returning to their baselines after stressors, whereas males went lower than initial sEMG baselines, as also found by Krantz et al. (2004). Our clinical experience is that males are often more aware, and more capable, of reducing shoulder tension. There was a trend of females displaying more sEMG reactivity during all stressors, with significance reached during the preprogram emotional stressor and the postprogram cognitive stressor. Krantz et al. used different stressors and found that females had more muscle response for the Stroop test but not arithmetic or cold pressor stressors. Because most studies have grouped males and females together, it is important for future studies to look at male-female differences and a variety of stressors to identify differential responses.

Although the females in this study had very low sEMG levels, they had a positive correlation between higher anxiety and higher muscle tension levels, and they also reported significantly higher symptoms and medication use. It is possible that the sEMG reflects increased breathing, because Masaoka and Homma (2001) showed that an increased respiratory rate is positively correlated with STAI trait anxiety scores.

Arena and Hobbs (1995) cited early research that associated high anxiety with increased physiological responses. Krantz et al. (2004) reported that sEMG and sympathetic nervous system activity are related and may be responsible for higher anxiety and more stress-related problems, whereas Izard (1990) reported a relationship between emotions and facial EMG. Conrad and Roth’s review article (2007) noted higher muscle tension in those with anxiety, but males and females were grouped together. A larger sample size comparing males and females is needed to answer the question of whether there are real male-female differences in the relationship between sEMG and anxiety.

Subjective Versus Objective Findings

Like others (Ford et al., 1983; Hoehn-Saric et al., 1997; Hoareau et al., 2021) who have reported that physical measures fail to correlate with psychological measures, this study’s participants reported subjective improvement in well-being at much higher levels than were found in any measured physiological changes between the pre and post stress profiles. Self-report measures of emotion and emotion regulation may not always cohere with psychophysiology (Baum & Grunberg, 1991; Burr et al., 2021). Even more interesting is that when confronted with the mismatch between perception and actual psychophysiological behavior, clients express disbelief (Schilling & Poppin, 1983).

Burr et al. (2021) also reported that anxiety predicts how individuals regulate in the laboratory, but how they report regulating in the real world is not highly correlated with how they regulate in the laboratory. This calls for future research comparing psychological and physiological measures from both the laboratory and real-world assessments. Ideally, the laboratory stressors will be more similar to real-life activity, perhaps using virtual reality, and will have more assessment conducted in the real world, perhaps using wearable devices and artificial intelligence for data collection.

Male-Female Differences

Females consistently demonstrated significantly colder hands and higher heart rate and a pattern of higher sEMG than males did. Perhaps this is due to genetics (Bohannon, 2023), behaviors (Cohen & Lansing, 2021; Taylor et al., 2000), small sample size, variability, or other factors. Research shows there are male-female differences in physiological and psychological parameters related to stress responses and their management, and these should be reflected in programming. One possible contributor to such differences is Dolcos et al.’s (2020) finding that females have enhanced sensitivity to emotional stimuli, especially negative stimuli, and they show negative affective bias in attention and perception while also exhibiting enhanced emotional competence in the processing of emotions. In addition, Rattel et al. (2020) noted females have a higher/stronger concordance (response concordance is the integration of thinking, feeling, and responding) than males on almost all physiological parameters and that females are better than males at recognizing emotions, expressing emotions, and being more aware of emotions. Rattel et al. further postulated that these sex differences may be due to better awareness as women are more responsive and sensitive to the environment.