Viability of resting electroencephalograph asymmetry as a predictor of exercise-induced affect: a lack of consistent support.

Lochbaum, Marc R.

Acute exercise is a highly effective mood-regulating strategy when compared to passive stimulation strategies such as drinking coffee and watching television (Thayer, Newman, & McClain, 1994). This mood-regulating strategy is most likely a result of the reductions in anxiety and improvements in positive mood/energy typically reported in the literature (Landers & Arent, 2001). Many theories and perspectives have been investigated such as self-determination theory (Lutz, Lochbaum, & Tumbow, 2002), Solomon's opponent-process theory of acquired motivation (Lochbaum, Karoly, & Landers, 2004), and personality and goal cognitions (Lochbaum and Lutz, 2005) to explain this relationship. To date no one theory or perspective has been dominant. Researchers, in addition to testing theory, have investigated psychological and physiological mechanisms to explain the exercise-affect relationship. One physiological mechanism that has demonstrated some promise in the published literature is that of resting asymmetry activation, specifically in the frontal region (Petruzzello, Hall, & Ekkekakis, 2001; Petruzzello & Landers, 1994; Petruzzello & Tate, 1997).

As proposed by Davidson (1994, 1998), resting asymmetry is a biological marker of one's propensity to approach or withdraw from a situation. Approach behaviors are associated with positive affect such as happiness, whereas withdrawal behaviors are associated with negative affect such as anxiety (Davidson, 1993a). Davidson (1993b) has proposed that these emotional reactions to a sufficient emotion eliciting stimulus are superimposed on asymmetrical frontal activation and represented by relative right versus left resting frontal differences that are typically measured by electroencephalography (EEG) in the alpha band of normal brain activity (i.e., 8 to 13 Hz). To this end, Davidson and his colleagues (e.g., Davidson, 1984; Davidson, Ekman, Saron, Senulis, & Friesen, 1990; Davidson & Tomarken, 1989) have demonstrated that this individual difference in emotional reactivity is predictive of future affective response (Wheeler, Davidson, & Tomarken, 1993) and concurrent affect (Henriques and Davidson, 1990), though opposition to their work has been reported (Hagemann, Naumann, Becker, Maier, & Bartussek, 1998; Reid, Duke, &Allen, 1998).

In the domain of exercise psychology, several studies have examined the predictive ability of resting frontal EEG asymmetry in an attempt to understand affective change in response to acute bouts of aerobic exercise (Hall, Ekkekakis, Van Landuyt, & Petruzzello, 2000; Petruzzello et al., 2001; Petruzzello & Landers, 1994; Petruzzello & Tate, 1997). As a predictor of exercise induced affect, resting frontal EEG asymmetry has significantly accounted for 18% to 30% of postexercise anxiety (Petruzzello & Landers, 1994; Petruzzello & Tate, 1997) and 6.4% to 22.7% of positive affect/energy (Petruzzello et al., 2001) following 30 minute runs at approximately 70 to 75% of maximal oxygen consumption. It is important to note that while the above mentioned investigations provide support for the predictive ability of resting frontal EEG asymmetry, one study provided no support for this mechanism (Hall et al., 2000). In addition, within the above mentioned investigations, resting frontal EEG asymmetry was predictive of affect and anxiety measures in only 7.8% of the postexercise measurements (positive affect/ energy: 15.3%; negative affect/tension: 0.0%; state anxiety: 15.0%).

Given that the asymmetry index has not consistently accounted for significant variance in post exercise affect, it is possible to question why another investigation is warranted. The present investigation is warranted based on two reasons. Past investigations have contained a severe methodological failure as well as highlighting the importance of fitness level. Concerning the methodological failure, Wheeler, Davidson, and Tomarken (1993) stated that "frontal asymmetry culminates in an emotional response only in conjunction with the requisite environmental input" (p. 83). In the absence of environmental input, individuals with asymmetrical frontal activation may not display positive or negative affective reaction. Unfortunately, much of the past research has not demonstrated significant affect or anxiety changes. For example, Petruzzello and Tate (1997) reported that from pre- to postexercise few affective changes occurred. Hall and colleagues (2000) reported much the same for tense arousal and state anxiety. Hence, it appears that the basic requirement for testing of the asymmetry index has not been consistently met. This limitation has been discussed in the literature (Petruzzello & Tate, 1997). To address this problem, the present study measured affective change both during and post exercise. It was hypothesized that affect and anxiety changed during exercise would be significantly different from pre-exercise values based on past research (Bixby, Spalding, & Hatfield, 2001; Petruzzello, Jones, & Tate, 1997) and therefore provide a test of the asymmetry index even if pre- to postexercise changes were not significant.

Though this methodological flaw exists in the literature, Petruzzello and colleagues (2001) have demonstrated the following interesting and unique finding: when interacted with fitness level, the asymmetric index becomes a more powerful predictor of postexercise affect. More specifically, Petruzzello and colleagues (2001) reported that the asymmetric index predicted significant amounts of variance (6.7% to 11.5%) in the energy component of energetic arousal, whereas it did not for the low fit individuals. To investigate this unique finding, the present investigation sought out participants with very different aerobic activity histories to ensure fitness differences. Hence, the overall purpose of this investigation was to examine whether Davidson's (1992) resting frontal asymmetric index is a viable mechanism to explain exercise induced affective change with special attention being paid to affect and anxiety being measured during exercise and the participants' aerobic fitness histories.

Method

Participants

Participants were 53 volunteer, university students (28 active: 15 male, 13 female; 25 inactive: 13 men, 12 women). All participants were recruited via advertisement and personal communications from exercise science and psychology courses at a large southwestern university. Participants were required to meet stringent requirements for being classified as either as an aerobically active or inactive to help ensure fitness level differences. Participants completed the Measurement of Habitual Physical Activity questionnaire (Baecke, Burema, & Frijter, 1982) as the screening tool. The criteria for being classified as an active or trained exerciser involved the frequency, duration, and intensity of the exercise and the duration of training. The American College of Sports Medicine (2000) has recommended 3 to 5 days as the frequency and 20 to 60 minutes as the duration. Intensity requirements vary depending on the duration of training. Pollock and Wilmore (1990) suggested that training differences might occur after a minimal time (e.g., 8 weeks) and duration (e.g., 15 min), but training for a longer duration (> 30 min and > 20 weeks) results in greater aerobic fitness changes. Hence, to be included as an active exerciser in the present study, participants were required to have exercised at least three times per week for 45 min at a moderate intensity over the last 6 months.

The criteria for being considered untrained or sedentary were based on detraining data found in Pollock and Wilmore (1990). Because cardiovascular inactivity of greater than 8 weeks ensures that the participants' [VO.sub.2max] will be similar to their typical [VO.sub.2max] even if they had been engaged in cardiovascular training prior to current inactivity, participants in the present study were required to have no cardiovascular or other type of fitness training for the 6 months prior to their participation. Hence, the inactive participants had to have reported no leisure time physical participation over the 6 months prior to the study.

Measures

Self-reported anxiety. The short form version of the state anxiety inventory (Form Y-l,) was used (Spielberger, 1979). This inventory consists of 10 items that are identical to half of the items used in the full 20-item inventory (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983). This short form is suited for studies with repeated measure designs where the repeated assessments need to be made quickly. The full 20-item inventory has been shown to be a valid measure of anxiety (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983). Many lines of research providing evidence for concurrent, divergent, convergent, and construct validity of the 20-item inventory have been reported (Spielberger, et al., 1983). The shortened version as reported in Spielberger (1979) is highly related to the 20-item scale in that the correlations between the two measures are extremely high (e.g., for Navy recruits: 198 males, r= .94; 72 females, r= .96 and for college students: 66 males, t= .95; 133 females, r= .95). Participants responded to the statement "how do you feel currently?" when answering each question.

Self-reported affect. The Activation Deactivation Adjective Checklist (ADACL; Thayer, 1989) was used to assess affect. The ADACL is a 20-item self-report inventory that assesses energetic arousal and tense arousal that are consistent, respectively, with dimensions of positive activation (positive affect) and negative activation (negative affect) in Watson, Wiese, Vaidya, and Tellegen's (1999) model. The ADACL was chosen over other measures of affect for several important reasons. First the ADACL is a theoretically-based modal of activation that is relevant in an exercise setting. The problems of measures that do not incorporate activation in an exercise context have been reviewed by Ekkekakis, Hall, and Petruzzello (1999). Second, the ADACL provides a representation of the global affective space. Finally, the ADACL's reliability and construct validity are well established (Thayer, 1986). In the present sample, the internal reliabilities across all measurement time points for energetic arousal and tense arousal ranged from .72 to .90.

Physical activity level. The Measurement of Habitual Physical Activity questionnaire (Baecke, Burema, & Frijter, 1982) was given to the active participants in order to screen and then classify participants as either aerobically active or inactive. The participants were asked to report their history of aerobic exercise (years of training, and frequency, intensity, and duration of each training bout) and their history of any other form of exercise involvement.

Brain Activation. Brain activation was assessed at the frontal (F3, F4) sites from both hemispheres. To record this measure, a lycra electrode cap (Electro-Cap, Inc., Eaton, OH) was fitted on the subject's head adjusted so that the distance from the front of the cap to the bridge of the nose was equivalent to one-tenth of the distance from the protrudence at the base of the scalp to the bridge of the nose. The electrode caps come in four sizes (46-50 cm, 50-54 cm, 54-58 cm, 58-62 cm) to ensure proper relative electrode placement (Blom & Anneveldt, 1982) and they are made of elastic spandex-type fabric with recessed, pure tin electrodes sewn in. Electrode gel was applied to the electrodes of interest to create the conductivity needed for taking scalp measures of EEG activity. Electrical impedance was measured at 30 Hz and electrode gel was reapplied to any EEG sites that had impedance greater than 5 K ohms. All leads were referenced to the left and right linked mastoids and a ground electrode in the lycra electrode cap. Ocular artifact was assessed by electro-oculogram recording from the supra-orbit and external canthus of the subject's left eye. EEG was collected using a Grass Model 12 Neurodata Acquisition System physiograph with software developed by Neuroscan, Inc. The high and low bandpass filters for the EEG measurement were set at 0.1 Hz and 100 Hz respectively and the signals were amplified 50,000 times. The high and low bandpass filters for the electro-oculogram measures were set at 3 Hz and 100 Hz respectively and the signals were amplified 20,000 times. The sampling rate for all signals was 256 Hz. Participants were required to sit quietly in an uptight position with their eyes closed. The baseline data were collected 10 min prior to the start of the conditions in eight consecutive 60 s trials.

The EEG waveforms were visually inspected for artifact. Artifact was removed from each EEG trial across all channels prior to further data analysis. In addition, the artifact-removed data were subjected to an electro-oculogram artifact reduction procedure to ensure removal of all ocular movement artifacts. Artifact-free data were then subjected to a fast Fourier transformation for decomposition of the EEG waveform into sine wave components. The components were used to estimate spectral power, which was converted to a power density function as a measure of the average spectral power across the 8-minute trials. Next, a natural log transformation was applied to the EEG data to normalize distributions. From this data, an EEG asymmetry index was derived (i.e., log right hemisphere--log left hemisphere alpha power). The data reduction methods just described replicated those used by previous researchers (e.g., Tomarken, Davidson, Wheeler, & Kinney, 1994). Tomarken and colleagues (1992) have demonstrated excellent stability for resting EEG absolute power for described collection of frontal alpha activity.

Procedure

Participants were initially asked, via mass screening in undergraduate courses, whether they met the criteria for being aerobically active or inactive. Participants meeting the criteria were then called and asked to participate. As the data for this study were collected within a larger project, additional procedures were followed that are explained elsewhere (Lochbaum, Karoly, & Landers, 2002, 2004). Specific to this investigation, each participant was required to visit the research laboratory on three separate days. The first session required the participants to complete the Human Participants Consent Form and to perform a graded exercise protocol (Astrand & Rodahl, 1977) to determine maximal oxygen consumption. On the following test days, participants exercised for 30 continuous minutes at 70-75% and 50-55% of their [VO.sub.2max]. The order of these days was counterbalanced across all participants. Prior to each condition, EEG preparation and collection was completed as previously described. After completion of pre-exercise EEG recordings, participants completed the AD ACL and the short form SAI prior to, 5, 10, and 25 minutes during the sessions, and immediately, 10, and 20 minutes following the termination of the exercise sessions.

Results

Gender, Group, and Exercise Session Differences

Given the inclusion of both males and females in this study, the effects of gender were examined for the self-reported measures and EEG asymmetry index values. These preliminary analyses included gender as an independent variable within both multivariate analysis of variance and regression analyses. The results of these analyses indicated that significant differences did not exist, thus, gender was not included in further analyses. Hence, only group differences were initially examined. Effect sizes (ES) were calculated as deemed appropriate as indicates of the meaningfulness of differences (Hedges, 1981). As for group differences, age and height did not differ though weight, F(1,52) = 44.42,p < .05, ES = -.62, did as a function of group (active M = 67.39 [+ or -] 10.55 kg; inactive M = 80.00 [+ or -] 27.30 kg). As a confirmation of activity status, a significant main effect emerged for [VO.sub.2max], F(1, 52) = 44.42,p < .001, ES = 1.83, such that active participants (M = 49.92 [+ or -] 5.38 ml/kg/min) had a greater [VO.sub.2max] than inactive participants (M = 39.30 [+ or -] 6.21 ml/kg/min), thus, revealing expected and necessary fitness differences between the two groups for future analyses. Several analyses were conducted on both participant groups by the two exercise sessions to determine if the sessions were indeed different with regards to workload. In brief the sessions were significantly different. These results are reported in Lochbaum et al. (2004).

Affect and Anxiety Reporting Patterns

The primary requirement to test resting frontal asymmetry's ability to predict affect and anxiety is that the exercise stimulus resulted in significant affective and anxiety variation from baseline. Therefore, it was critical to determine whether affect and anxiety during and after exercise was significantly different from baseline values. To simplify this determination, the three measurement time points (see Table 1) during and after exercise were averaged for each session. Given that research (Ekkekakis, Hall, VanLanduyt, & Petruzzello, 2000; Felts & Vaccaro, 1988; Porcair, Ebbeling, Ward, Freedson, & Ripe, 1989) has demonstrated that low intensity exercise conditions elicit the same affective responses equal to higher intensity conditions (excluding maximal) no hypotheses were warranted; hence, a exercise condition factor was not examined. The analyses of most importance were the time main effects and the potential activity influences (interaction term). In order to assess these effects, a group (2: active, inactive) by time (2: resting, during exercise, postexercise) repeated measures MANOVA was conducted for each exercise intensity.

For 55% exercise session, the main effect for time, Wilks' [lambda] = .33, F(6, 46) = 15.34,p < .01, and group were significant, Wilks' [lambda] = .75, F(3, 49) = 5.21,p < .01, as was the 2-way Group by Time interaction, Wilks' [lambda] = .69, F(6, 46) = 3.32, p < .01. This repeated measures MANOVA was followed up by separate univariate ANOVA's with repeated measures on the time factor to confirm the significant time main effect and determine the exact causes for this main effect. In short, the time main effects for energetic arousal and tense arousal were highly significant (p < .001), whereas the time main effect for state anxiety was not significant. Paired-sample t-tests were conducted with alpha set at .01 to examine the time main effects for energetic arousal and tense arousal. Energetic arousal and tense arousal were significantly elevated from baseline both during and after the exercise session (ES's energetic arousal: 1.06 and 1.09; tense arousal: .48 and .72). The Group by Time interaction (see Figure 1) was significant (p [less than or equal to] .01) for both tense arousal and state anxiety, but not for energetic arousal. The interactions were due to the inactive group's significant increases in tense arousal and state anxiety over time; whereas the active participants' reported no significant changes in tense arousal over time and a significant decrease in state anxiety from pre- to postexercise. At both time points, the two groups differed significantly (p < .05).

[FIGURE 1 OMITTED]

For 70% exercise session, the main effect for time, Wilks' [less than or equal to] =. 10, F(6, 46) = 63.46,p < .01, and group were significant, Wilks' [less than or equal to] = .75, F(3, 49) = 5.44, p < .0 l, as was the 2-way Group by Time interaction, Wilks' [less than or equal to]. = .67, F(6, 46) = 3.72, p < .01. This repeated measures MANOVA was followed up by separate univariate ANOVA's with repeated measures on the time factor to confirm the significant time main effect and determine the exact causes for this main effect. In short, the time main effects for each variable were highly significant (p < .001). Paired-sample t-tests were conducted with alpha set at .01 to examine the time main effects. Energetic arousal and tense arousal were significantly elevated from baseline both during and after the exercisesession (ES's: energetic arousal: 1.83 and 1.09; tense arousal: 2.35 and .72), whereas state anxiety was only significantly different (elevated) during exercise, ES = 1.58. The Group by Time interaction (see Figure 2) was significant (p [less than or equal to] .0 l) for both energetic arousal and tense arousal, but not state anxiety. The significant interaction for tense arousal was due to the inactive groups heightened response during and after exercise when compared to the active group's tense arousal response. In addition, tense arousal scores at all time points differed significant within both participant groups as confirmed by paired t-tests (see Figure 2). For energetic arousal the significant interaction is more difficult to explain given the two groups did not differ significantly at any of the measurement time points as confirmed by one-way ANOVA's at each time point. All of the paired t-tests were significant and confirmed that energetic arousal responses within each participant group differed from one another (pre to during, during to post, and pre to post).

[FIGURE 2 OMITTED]

Resting EEG Asymmetry Predicting Affect and Anxiety

Prior to examining the predictive ability of resting frontal asymmetry during and postexercise, a one-way ANOVA confirmed that no statistically significant differences existed in asymmetry index between the active (M [+ or -] SEM asymmetry = .003 [+ or -] .025, 55% intensity; .072 [+ or -] .026, 70% intensity) and inactive (M [+ or -] SEM asymmetry = .032 [+ or -] .034, 55% intensity; .051 [+ or -] .038, 70% intensity) prior to the either exercise intensity. To determine the predictive power of resting EEG asymmetry over and above that of baseline affective state, a set of hierarchical regression analyses were performed for each measured affective variable at all measurement time points.

Resting EEG asymmetry accounted for no unique variance in energetic arousal, tense arousal, or state anxiety during or after exercise in the 55% condition. For the 70% condition (see Table 2), resting EEG asymmetry significantly (p [less than or equal to] .05) accounted for 9.1% of tense arousal at 5 minutes into the exercise session. The asymmetry index also accounted for 6.7% of tense arousal 25 minutes into the 70% exercise condition and 6.1% of state anxiety at 5 minutes into the exercise. These results only approached significance (p [less than or equal to] .075).

Resting EEG Asymmetry and Physical Activity History Predicting Affect and Anxiety

To examine the second main purpose of the present investigation, activity level was examined in a second set of hierarchical regression analyses to determine whether it interacted with the asymmetry index. For these hierarchical regression analyses, baseline affect was first entered followed by resting frontal asymmetry, activity level, and the interaction term. Significant activity level by frontal asymmetry interactions were found for energetic arousal at 5, 15, and 25 minutes during the 55% exercise condition accounting for 6.3%, 7.4%, and 9.4% of additional variance, respectively. In the 70% condition, the interaction term was significant and accounted for an additional 9.3% of variance for energetic arousal at 15 minutes during the exercise session. In addition, significant interactions emerged for tense arousal at 15 minutes into the 55% and 70% conditions accounting for 4.7% and 5.8% of additional variance in the tense arousal score, respectively. The interaction term was not statistically significant for any of the measures collected following either exercise session. The significant findings during both exercise conditions were followed up by performing separate regressions analyses for each activity group to determine the importance of each participant group upon the affective scores.

The separate regression analyses (see Table 3) for the active group revealed the predictive power of the resting frontal asymmetry index to be 18.4%, 19.1%, and 23.3% over and above that of baseline energetic arousal at 5, 15, and 25 minutes during the 55% exercise condition, respectively. In addition, for the active group, resting frontal asymmetry accounted for 16.4% of variance over and above that of baseline tense arousal 15 minutes into the 55% exercise condition. For the inactive group, resting frontal asymmetry accounted for 26.3% of additional variance in tense arousal 15 minutes into the 70% exercise condition. Last, given that the regression analyses included activity status as a dummy coded variable prior to the interaction term, it is important to gain an understanding of those effects. As can be seen in Table 4, activity status was a significant predictor of affect and anxiety response in numerous instances and accounted in one instance for 34.6% of post exercise state anxiety. These highly significant findings suggest that the activity status of our participants greatly influenced affective reporting in the present data.

Discussion

The purpose of the present investigation was to determine whether resting brain activation is a viable explanation for the relationship between acute aerobic exercise and affect or anxiety. To best test this purpose, affect and anxiety were measured both during and after exercise to increase the chances of measuring sufficient environmental input in two bouts of aerobic exercise of different intensities. In addition, given that past research has demonstrated that fitness level interacts with the resting brain activation (Petruzzello et al., 2001), two very different groups of both male and female participants based on six month exercise history were recruited for the present investigation. It was found that resting frontal EEG asymmetry accounted for no unique variance in our dependent measures in the 55% aerobic exercise condition and in only one instance for tense arousal in the 70% condition.

Based on Davidson's (1992) framework, these findings are disappointing given the fact that affect and anxiety differed significantly from pre-exercise affect and anxiety in nearly every instance. These significant differences satisfied Davidson's requirement that a significant stimulus (the exercise sessions) was experienced. It appears that by itself, resting frontal EEG asymmetry is not a reasonable mechanism to explain the exercise-affect or anxiety relationship. Taken as a whole, resting frontal EEG asymmetry has not explained any unique variance in energetic arousal, tense arousal, or state anxiety in the published literature (Hall et al., 2000; Petruzzello & Tate, 1997) or in the present investigation in low intensity conditions ([less than or equal to] 55% of maximal oxygen consumption). In higher intensity conditions, unique variance in tense arousal has been explained by the asymmetry index only in the present investigation. Resting frontal EEG asymmetry has explained unique variance in energetic arousal and state anxiety in higher intensity conditions (Petruzzello et al., 2001). It is important to remember that not all of published literature has reported consistent differences in pre to postexercise affect and anxiety that could explain for a lack of significant findings. Of those that have, very few instances exist whereby the asymmetry index has explained significant and unique variance in postexercise affect and anxiety.

Interestingly, in the present investigation, when the asymmetry index accounted for unique variance in tense arousal and state anxiety, the difference between tense arousal and state anxiety measured during exercise and their baseline scores were significant and very meaningful (ES's > 1.50). This predictive ability was not consistent given the other differences from baseline to during exercise were both significant and meaningful (ES's > 1.00). In addition, postexercise values also differed significantly and meaningfully in most instances from preexercise values; but, the asymmetry index failed to account for any unique variance. Again, based on Davidson's framework, the organism must be challenged for any relationship between the index and affect to be expected. It could be that the index will only be predictive during exercise when the organism is being challenged as opposed to postexercise when the organism is "cooling down" from the challenge. This challenge is demonstrated by such large and meaningful differences as represented by the effect size parameter. In retrospect, the published literature and the present investigation are measuring affect and anxiety 30 minutes after the challenging stimulus has been terminated. Hence, in addition to the challenge being eliminated, participants perhaps have moved onto other mental thoughts dramatically reducing the salience of affect or anxiety as a result of physical exercise.

The second portion of this investigation examined the importance of aerobic fitness by recruiting participants with very different physical activity histories to the prediction of affect and anxiety. To examine this importance, a dummy coded variable was created using the two participant groups and their corresponding asymmetry index entered into the hierarchical regression analyses. These analyses partially supported past research (Petruzzello et al., 2001) in that the interaction term was predictive of positive affect/energy for the active or fit participants. The major difference is that Petrnzzello et al. (2001) demonstrated this predictive ability in a higher intensity condition post exercise, whereas for the present investigation positive affect or energetic arousal was predicted in a lower intensity condition during the exercise session. In addition and in contrast to past research, the interaction term was significant for the low fit or inactive participants for negative affect or tense arousal 15 minutes during both the low and high intensity conditions. Although the findings clearly demonstrate that the asymmetry index significantly interacted with fitness predictive in several instances, the pattern of findings was inconsistent given that significant and meaningful differences occurred in nearly all baseline to during and after exercise affect and anxiety values in the higher intensity condition. Hence, no one psychological or physiological reason stands out that would explain why the index interacting with fitness would be predictive mostly during the low intensity condition. This is due to the fact that that several researchers have demonstrated that low and high intensity conditions produce similar affective responses (Ekkekakis, et al., 2000; Felts & Vaccaro, 1988; Porcair et al., 1989).

Prior to completely dismissing the asymmetry index as a viable explanation for affective responses to physical exercise, it is worth mentioning that unique variance in affect and anxiety were significantly explained in most instances by the dummy coded activity or fitness status variable. These results were most prominent for negative affect and anxiety; therefore, a significant portion of variance was already explained. Past research is somewhat mixed as to the importance of activity and or fitness level on affective and anxiety response to acute aerobic exercise (e.g., Blanchard, Rogers, Spence, & Coumeya, 2001; Steptoe, Kearsley, & Waiters, 1992). The present data clearly demonstrate the importance of participant activity history and thus current fitness level upon affective and anxiety responses (1). In summary, the purpose of the present investigation was to determine the viability of Davidson's asymmetry index as a predictor of affective and anxiety response in response to aerobic exercise. Taken as a whole, the present investigation and past investigations offer inconsistent support for the asymmetry index as a viable explanation for affective and anxiety responses to acute bouts of lower and higher intensity aerobic exercise.

References

American College of Sports Medicine (2000). Guidelines for exercise testing and prescription (6th Ed.). Philadelphia: Williams & Wilkins.

Astrand, P., & Rodahl, K. (1977). Textbook of work physiology: Physiological bases of exercise (pp. 331-361). NY: McGraw-Hill.

Baecke, J. A. H., Burema, J., & Frijters, J. E. R. (1982). A short questionnaire for the measurement of habitual physical activity in epidemiological studies. American Journal of Clinical Nutrition, 36, 936-942.

Bixby, W. R., Spalding, T. W., & Hatfield, B. D. (2001). Temporal dynamics and dimensional specificity of the affective response to exercise and varying intensity: Differing path ways to a common outcome. Journal of Sport and Exercise Psychology, 23, 171-190.

Blanchard, C. M., Rodgers, W. M., Spence, J., & Courneya, K. S. (2001). Feeling state response to acute exercise of high and low intensity. Journal of Science and Medicine in Sport, 4, 30-38.

Blom, J. L., & Anneveldt, M. (1982). An electrode cap tested. Electroencephalography and Clinical Neurophysiology, 54, 591-594.

Davidson, R.J. (1984). Affect, cognition, and hemispheric specialization. In C.E. Izard, J. Kagan, & R. Zajonc (Eds.), Emotions, cognition, and behavior (pp. 320-365). New York: Cambridge University Press.

Davidson, R. J. (1994). Asymmetric brain function, affective style and psychopathology: The role of early experience and plasticity. Development and Psycholopathology, 6, 741-758.

Davidson, R. J. (1993a). The neuropsychology of emotion and affective style. In M. Lewis & J. M. Haviland (Eds.), Handbook of Emotions (pp. 143-154). New York: Guilford.

Davidson, R. J. (1993b). Parsing affective space: Perspectives from neuropsychology and psychophysiology. Neuropsychology, 7, 464-475.

Davidson, R. J. (1998). Anterior electrophysiological asymmetries, emotion, and depression: Conceptual and methodological conundrums. Psychophysiology, 35, 607-614.

Davidson, R.J., Ekman, P., Saron, C.D., Senulis, J., & Friesen, W. (1990). Approach/withdrawal and cerebral asymmetry: Emotional expression and brain physiology I. Journal of Personality and Social Psychology, 58, 330-341.

Davidson, R.J., & Tomarken, A.J. (1989). Laterality and emotion: An electrophysiological approach. In F. Boiler & J. Grafman (Eds.), Handbook of neuropsychoiogy (pp. 419-441). New York: Elsevier.

Ekkekakis, P., Hall, E. E., VanLunduyt, L. M., & Petruzzello, S. J. (2000). Walking in (affective) circles: Can short walks enhance affect? Journal of Behavioral Medicine, 23, 245-275.

Felts, W. M., & Vaccaro, P. (1988). The effect of aerobic exercise on postexercise state anxiety and psychological arousal as a function of fitness level. Clinical Kinesiology, 42, 89-96.

Hagemann, D., Naumann, E., Becker, G., Maier, & Bartussek, D. (1998). Frontal brain asymmetry and affective style: A conceptual replication. Psychophysiology, 35, 372-388.

Hall, E. E., Ekkekakis, P., Van Landuyt, L. M., & Petruzzello, S. J. (2000). Resting frontal asymmetry predicts self-selected walking speed but not affective responses to a short walk. Research Quarterly for Exercise and Sport, 71, 74-79.

Henriques, J.B., & Davidson, R.J. (1990). Regional brain electrical asymmetries discriminate between previously depressed subjects and healthy controls. Journal of Abnormal Psychology, 199, 22-31.

Hedges, L. V. (1981). Distribution theory for Glass's estimator of effect size and related estimators. Journal of Educational Statistics, 6, 107-128.

Landers, D. M. & Arent, S. (2001). Physical activity and mental health. In R. Singer, H. Hausenblas, & C. M. Janelle (Eds.), Handbook of Sport Psychology (pp. 740-765). New York: John Wiley & Sons, Inc.

Lochbaum, M. R., Karoly, P., & Landers, D. M. (2002). Evidence for the importance of openness to experience on performance of a fluid intelligence task by physically active and inactive participants. Research Quarterly for Exercise and Sport, 73,437-444.

Lochbaum, M. R., Karoly, P., & Landers, D. M. (2004). Affect responses to acute bouts of aerobic exercise: A test of opponent-process theory. Journal of Sport Behavior, 27, 330-348.

Lochbaum, M. R., & Lutz, R. (2005). Exercise enjoyment and psychological response to acute exercise: The role of personality and goal cognitions. Psychology Research Journal, 1, 4-12.

Lutz, R., & Lochbaum, M. R., & Turnbow, K. (2003). The role of autonomy in post-exercise affect responding. Journal of Sport Behavior, 26, 137-154.

Petruzzello, S. J., Hall, E. E., & Ekkekakis, P. (2001). Regional brain activation as a biological marker of affective responsivity to acute exercise: Influence of fitness. Psychophysiology, 38, 99-106.

Petruzzello, S. J., Jones, A. C., & Tate, A. K. (1997). Affective responses to acute exercise: A test of opponent-process theory. Journal of Sports Medicine and Physical Fitness, 37,205-212.

Petruzzello, S.J., Landers, D.M., Hatfield, B.D., Kubitz, K.A., & Salazar, W. (1991). A meta-analysis on the anxiety reducing effects of acute and chronic exercise. Sports Medicine, 11,143-182.

Petruzzello, S.J., & Landers, D.M. (1994). State anxiety reduction and exercise: Does hemispheric activation reflect such change? Medicine and Science in Sports and Exercise, 26, 1028-1035.

Petruzzello, S. J., & Tate, A. K. (1997). Brain activation, affect, and aerobic exercise: An examination of both state-independent and state-dependent relationships. Psychophysiology, 34, 527-533.

Pollock, M. L., & Wilmore, J. H. (1990). Exercise in health and disease. Philadelphia: W. B. Saunders Company.

Reid, S. A., Duke, L. M., Alien, J. J. B. (1998). Resting frontal electroencephalographic asymmetry in depression: Inconsistencies suggest the need to identify mediating factors. Psychophysiology, 35, 389-404.

Thayer, R. E. (1986). Activation-deactivation adjective check list: Current overview and structural analysis. Psychological Reports, 8, 607-614.

Thayer, R. E. (1989). The biopsychology of mood and arousal. Oxford: Oxford Press.

Thayer, R.E., Newman, R., & McClain, T.M. (1994). Self-regulating of mood: Strategies for changing a bad mood, raising energy, and reducing tension. Journal of Personality and Social Psychology, 67, 910-925.

Tomarken, A.J., & Davidson, R.J. (1994). Frontal brain activation in repressors and nonrepressors. Journal of Abnormal Psychology, 103, 339-349.

Tomarken, A.J., Davidson, R.J., Wheeler, R.E., & Kinney (1992). Psychometric properties of resting anterior EEG asymmetry: Temporal stability and internal consistency. Psychophysiology, 29, 576-592.

Steptoe, A., Kearsley, N., & Waiters, N. (1992). Acute mood responses to maximal and submaximal exercise in active and inactive men. Psychology and Health, 8, 89-99.

Spielberger, C. D. (1979). Preliminary manual for the State-Trait Personality Inventory. Tampa, FL: Author.

Spielberger, C.D., Gorsuch, R.L., Lushene, R., Vagg, P.R., & Jacobs, G.A. (1983). Manual for the State-Trait Anxiety Inventory: STAI (Form Y). Palo Alto, CA: Consulting Psychologists.

Watson, D., Wiese, D., Vaidya, J., & Tellegen, A. (1999). The two general activation systems of affect: Structural findings, evolutionary considerations, and psychobiological evidence. Journal of Personality and Social Psychology, 76, 820-838.

Wheeler, R.E., Davidson, R.J., & Tomarken, A.J. (1993). Frontal brain asymmetry and emotional reactivity: A biological substrate of affective style. Psychophysiology, 30, 82-89.

Marc R. Lochbaum

Texas Tech University

Address Correspondence To: Marc Lochbaum, Ph.D., Health, Exercise, and Sport Sciences, Box 43011, Texas Tech University, Lubbock, TX 79409-3011, Phone: (806) 742-3371, E-mail: marc.lochbaum@ttu.edu

(1) Though the two participant groups differed significantly on [VO.sub.2max], all of the regression analyses were conducted with [VO.sub.2max] entered instead of the dummy coded physical activity variable to be sure that the pattern of results did not change. The results did not change in that the asymmetry index again was a very inconsistent predictor of affect and anxiety and [VO.sub.2max] was, as was activity status, a significant predictor of affect and anxiety in many instances.

Table 1. Means, standards deviations, and ranges for energetic arousal,
tense arousal, and state anxiety for all participants

 M SD Range

Energetic Arousal
 Pre 19.56 4.72 12.00-34.00
 Immediate post 20.74 3.75 12.00-27.00
 Post 10 18.83 3.92 10.00-28.00
 Post 20 18.61 3.78 10.00-26.00
Tense Arousal
 Pre 16.49 3.18 11.00-25.00
 Immediate post 20.74 3.75 12.00-27.00
 Post 10 18.83 3.92 10.00-28.00
 Post 20 18.61 3.78 10.00-26.00
State Anxiety
 Pre 15.05 2.36 10.00-19.00
 Immediate post 16.48 3.73 10.00-29.00
 Post 10 15.60 3.71 10.00-29.00
 Post 20 15.27 3.25 10.00-26.00

Table 2. Results of hierarchical regression analyses using resting
(pre-exercise) frontal asymmetry scores as a predictor of tense arousal
(TA) and state anxiety (SA) in the 70% exercise condition

 [R.sup.2]
Dependent variable Predictor Beta [R.sup.2] change

Tense arousal
 5 min during Pre-exercise TA .269 .069 .069
 Frontal asymmetry -.301 .160 .091
 25 min during Pre-exercise TA .145 .019 .019
 Frontal asymmetry -.259 .086 .067
State anxiety
 5 min during Pre-exercise SA .171 .021 .021
 Frontal asymmetry -.248 .082 .061

Dependent variable Predictor F change p

Tense arousal
 5 min during Pre-exercise TA 3.76 .058
 Frontal asymmetry 5.40 .024
 25 min during Pre-exercise TA 1.00 .321
 Frontal asymmetry 3.65 .062
State anxiety
 5 min during Pre-exercise SA 1.08 .303
 Frontal asymmetry 3.31 .075

Table 3. Results of separate hierarchical regression analyses for
each participant group

Dependent variable Predictor Beta [R.sup.2]

Active Participants
 Energetic arousal (EA)
 55% condition
 5 min during Pre-exercise EA .211 .022
 Frontal asymmetry .433 .205
 15 min during Pre-exercise EA .240 .031
 Frontal asymmetry .442 .222
 25 min during Pre-exercise EA .075 .000
 Frontal asymmetry .488 .233
 Tense arousal (TA)
 55% condition
 15 min during Pre-exercise TA .188 .037
 Frontal asymmetry -.405 .201
Inactive Participants
 70% condition
 15 min during Pre-exercise TA .007 .009
 Frontal asymmetry -.520 .272

 [R.sup.2]
Dependent variable change F change p

Active Participants
 Energetic arousal (EA)
 55% condition
 5 min during .022 .578 .454
 .184 5.77 .024
 15 min during .031 .824 .372
 .191 6.13 .020
 25 min during .000 .000 .987
 .233 7.58 .011
 Tense arousal (TA)
 55% condition
 15 min during .037 1.00 .325
 .164 5.12 .033
Inactive Participants
 70% condition
 15 min during .009 .204 .656
 .263 7.94 .010

Table 4. Results of hierarchical regression analyses for activity
status as a predictor of energetic arousal, tense arousal, and state
anxiety in both exercise conditions

 [R.sup.2]
Dependent variable F change change p

Energetic arousal
 70% condition
 25 min during 12.92 .180 .001
 Immediate post 3.61 .065 .063
 Post 10 min 7.98 .117 .007
Tense arousal
 55% condition
 5 min during 7.60 .130 .008
 15 min during 11.56 .047 .090
 25 min during 11.06 .181 .002
 Immediate post 1626 .225 <.001
 Post 10 min 7.67 .123 .008
 Post 20 min 4.85 .080 .032
 70% condition
 15 min during 2.94 .048 .093
 25 min during 9.49 .157 .003
 Immediate post 6.92 .109 .011
State anxiety
 55% condition
 5 min during 10.20 .160 .002
 15 min during 14.39 .220 <.001
 25 min during 19.94 .279 <.001
 Immediate post 8.56 .143 .005
 Post 10 min 5.46 .094 .024
 70% condition
 5 min during 4.67 .080 .035
 15 min during 13.72 .208 .001
 25 min during 21.42 .302 <.001
 Immediate post 31.13 .346 <.001
 Post 10 min 19.32 .273 <.001
 Post 20 min 9.16 .157 .004