Introduction

Studies have shown that participation in regular exercise can ameliorate symptoms of stress, anxiety, and depression (Salmon 2001) and that this enhanced mood profile can extend further into analgesic effects during exercise (Dietrich and McDaniel 2004). In addition, it has been shown that cognitive processes such as reaction time (Davranche and McMorris 2009) and neural arousal (Davranche and Pichon 2005; Lambourne and Tomporowski 2010) can be positively affected by exercise, though the degree of change among these variables is dependent on the intensity, duration, and type of exercise (McMorris and Hale 2012). Increasing support has also emerged for the positive effects of acute exercise on executive functions such as task switching, selective attention, working memory capacity, and inhibitory control (Guiney and Machado 2013; Verburgh et al. 2013). Suggested mechanisms for such improvements include neurogenesis and synaptogenesis through increased production of brain derived neurotrophic factor (BDNF), as well as augmented neurotransmitter levels and effectiveness (Lojovich 2010). Peripheral mechanisms, including the release and transport of catecholamines (epinephrine and norepinephrine) under a certain intensity threshold have been shown to positively affect cognition during and after low and moderate intensity exercise (McMorris and Hale 2012). It has been proposed, however, that increased concentrations of these catecholamines can lead to more neural noise at higher exercise intensities, meaning that higher order brain functions can become compromised during very strenuous activity. Though these findings suggest that exercise reliably influences cognition in a positive way, research has failed to come to a consensus on the effect of different types, length, or intensity of exercise on higher order cognitive function.

Some theories (Humphreys and Revelle 1984; Kahneman 1973) predict that CNS arousal, from the stress of physical activity and exercise, positively affects the responsiveness of sensory systems to various environmental stimuli that are presented (Lambourne and Tomporowski 2010). This heightened arousal state acts to facilitate information processing and cognitive function (Audiffren et al. 2008; Tomporowski 2003). During competition, athletes are often required to make quick decisions while simultaneously performing physical tasks (Godefroy et al. 2002). From a performance perspective, an increased rate of information processing likely benefits skill execution and reaction time. A review by Lambourne and Tomporowski (2010) showed that the positive effect of exercise on cognitive function was independent of exercise type (fatiguing, steady state, or inverted-U shaped). Although, it was also mentioned that anecdotally, there are many instances where fatiguing exercise acts to create a decrement in cognitive function. Increasing support for the inverted-U concept has emerged with some research suggesting increased processing speed following moderate intensity exercise; however, lack of improvements in accuracy have been attributed to a lack of effective test selection to assess this domain (Etnier and Chang 2009; McMorris and Hale 2012). The relationship between exercise and neural arousal is not simple, and the definitive timepoint or intensity where fatigue outweighs any cognitive benefit has yet to be determined.

Critical flicker fusion (CFF) threshold is a method of assessing overall central nervous system (CNS) arousal and sensory sensitivity. It is a visual/sensory discrimination task in which subjects are asked to determine at which point a flickering light increases in frequency and is perceived to become a fused light (ascending trial), and when a fused light decreases frequency and is perceived as a flickering light (descending trial). The higher the frequency that the subject can distinguish flickering light from a fused light, the greater the CNS arousal and sensory sensitivity present at the time of the task (Lambourne et al. 2010). CFF is classically presented as the mean of the ascending and descending trials (Mtot); a higher Mtot value represents increased arousal in the cerebral cortex. Another way of presenting the CFF data is by taking the difference between the mean ascending and descending trials, or Mdi. This shows the response criterion of the subject, where a larger difference indicates a conservative, or cautious, approach and smaller differences a liberal, or risky, approach (Davranche and Pichon 2005). CFF has been used in psychophysiological (Cavalade et al. 2015; Clemente-Suárez et al. 2017) and pharmacological (Hindmarch 1982; Smith and Misiak 1976) studies, and somewhat recently has been introduced into exercise science research (Davranche and Audiffren 2004; Lambourne et al. 2010).

CFF has been used in various studies to show how a single bout of exercise can affect cognitive functioning. Studies specifically using CFF to determine the effect of exercise on cognitive function are equivocal. Those that have assessed CFF before and after short, fatiguing exercise have shown either no change (Godefroy et al. 2002) or an increase (Davranche and Pichon 2005; Presland et al. 2005) in Mtot. Similarly, studies using steady-state exercise have also found either no change (Loy and O’Connor 2016; Suvi et al. 2016) an increase (Lambourne et al. 2010) or even a decrease (Grego et al. 2005) in Mtot. It is important to make the distinction that very few studies have reported the Mdi values. Two studies provided this value after fatiguing exercise (VO2max tests); Davranche and Pichon (2005) showed no change, which did not corroborate the results of Godefroy et al. (2002), which showed a decrease. Notably, these two studies used different testing modalities (cycling and treadmill, respectively) which has been shown to affect the results of cognitive tests (Lambourne and Tomporowski 2010).

To the knowledge of the authors, there have been no studies using CFF that have reported both the Mtot and Mdi with two distinctly different types of exercise (short, fatiguing and longer, steady-state) within the same study or examined the effect of exercise intensity. Therefore, the purpose of the present study was to determine if there is a positive change in cortical neural arousal after a short but fatiguing maximal exercise test, and after longer steady-state exercise at three different perceptually regulated intensities. We hypothesized that the short, fatiguing exercise would lead to a decrease in Mdi and an increase in Mtot, whereas the longer, steady-state exercise bouts would elicit opposite results.

Methods

Subjects and study design

22 subjects (10 men, 12 women) with a mean age of 25 ± 6 years volunteered to participate in the current study. All classified themselves as recreational runners and engaged in physical activity on a regular basis. Informed consent was obtained from all individuals included in the study, prior to participation. Subjects were asked to come to the laboratory for five total visits. The first was an orientation to the lab, including a familiarization to all equipment, and a chance to fully describe the details of the experiment outlined in the informed consent, and completion of the AHA/ACSM Pre-participation screening questionnaire (Thompson et al. 2009). The second visit included a VO2max test, with CFF assessments taken prior to and immediately after the test. The third, fourth, and fifth visits were all perceptually regulated 30-min treadmill runs. Each was at a different prescribed rating of perceived exertion [RPE; (Borg 1970)] and each included CFF assessments immediately before and after the test.

Critical flicker fusion (CFF)

Measurements of cortical neural arousal were determined before and immediately after exercise in each session. Subjects were seated comfortably in a dark room in front of the CFF device (Lafayette Instruments Flicker Fusion System, Model 12021A). They were asked to position their heads in the viewing chamber of the device; this chamber has two lights, one for each eye, separated by 7 cm. The distance from stimulus to the eye was 38 cm at a 1.9° angle, and the inside of the viewing compartment was completely black to minimize any reflection. The emitted light was presented for the right and left eye simultaneously. Subjects were asked to perform three ascending and three descending trials. Ascending trials began at 12 Hz and increased in frequency (1 Hz·s−1) until the subject determined that the lights changed from flickering to a solid light. Decreasing trials started at 50 Hz and ended when the subject determined that the light began flashing and was no longer perceived as solid. The subjects indicated this change by pressing a button on a handheld wired remote. These starting frequencies (12 and 50 Hz) have previously been used by multiple research groups (Al-Nimer and Al-Kurashy 2007; Curran et al. 1990; Ghozlan and Widlöcher 1993). The average frequency of the three ascending and three descending trials was then determined to be Mtot for that subject. To determine the subjective judgment criterion (Mdi), the difference between the mean ascending and descending values was calculated. Subjects thoroughly practiced the CFF testing during the first visit to the lab, to become familiar with the device.

Metabolic variables

During the VO2max test, all expired gases were collected with a True One 2400 metabolic cart (Parvomedics, Sandy UT). Both flow and gas calibrations were performed before each test. 15-breath rolling averages were calculated after all data collections were complete, and the maximal value for each variable [oxygen consumption (VO2), carbon dioxide production (VCO2), respiratory exchange ratio (RER), and ventilation (VE)] is presented. The primary criterion to determine if tests were considered truly maximal was a plateau in O2 uptake, defined as less than 2.0 ml·kg−1·min−1 increase between the maximal and immediately preceding values, using 15 s averaging. Secondary criteria, used when a plateau was not present, included either (1) HR within 10 bpm of the age-predicted maximal value using the 206.9–(0.67 × age) equation or (2) when there was an RER of at least 1.10. RPE could not be used as a secondary criterion due to the nature of the SPV test.

Heart rate (HR)

HR was continuously recorded throughout all four experimental trials (VO2max test and three 30-min runs) with a Polar chest strap and coded watch (Polar Electro Oy, Kempele, Finland). Subjects were blinded to the display of their HR.

Detailed procedures

Session 1 On the first day, the subjects read and signed the informed consent form, along with the health-screening questionnaire. Height and weight measurements for each subject were obtained. ACSM risk stratification procedures were used, and all subjects were classified as “low risk”. They were then familiarized with the equipment, the CFF device and the treadmill. Both the ascending and descending trials of CFF were performed multiple times; then, they were asked to practice using the treadmill (TrackMaster TMX 428CP, Full Vision Inc., Newton, KS) by running at RPE levels of 11, 13, and 15.

Session 2 During this session, the subjects completed a maximal exercise test using the self-paced VO2max (SPV) test protocol originally designed by Mauger and Sculthorpe (2012). Before stepping on the treadmill, the subjects were fitted with a Polar heart rate monitor. The treadmill was set to an incline of 8% as in the previous studies using this protocol (Hanson et al. 2016; Scheadler and Devor 2015) and used set RPE levels of 11 (very light), 13 (light), 15 (somewhat hard), 17 (hard) and 20 (very hard) for 2 min each. The tests lasted 10 min exactly. For this test, and all subsequent testing, subjects were blinded to the display of the treadmill. Before and immediately after the VO2max test, CFF testing was completed. Subjects were then allowed back on the treadmill to fully recover at a comfortable walking pace.

Sessions 3–5 These were identical except the RPE level was different for each session. RPE was randomly determined on the day of testing as either 13, 15, or 17; the subjects were told the condition when they arrived each day. They were given a 5 min warm up period, and then were fitted with the heart rate monitor; after this, they were asked to begin the protocol. They were expected to run for 30 continuous minutes at the prescribed RPE for the day, and were allowed to self-pace and adjust the treadmill speed throughout the entirety of each run, to maintain the RPE level. Before and after the run, CFF testing was performed.

Statistical analysis

IBM SPSS Statistics for Windows (Version 24, Armonk, NY) was used for all analyses and descriptive statistics. 4 (intensity: SPV, RPE13, 15 and 17) × 2 (time: pre/post) factorial repeated-measures ANOVAs were used to determine the effect of intensity condition and time point on cortical neural arousal using both Mtot and Mdi. Bonferroni corrections were used for post-hoc comparisons, and a Greenhouse-Geiser correction was used when the assumption of sphericity was violated. The alpha level was established a priori at p < 0.05 for all analysis.

Results

Full results from the SPV testing can be found in Table 1. The average HR for the three longer, steady-state conditions (RPE13, 15, and 17) were: 157.8 ± 17.8, 167.5 ± 17.4, and 175.7 ± 12.3 bpm, respectively. The 4 × 2 repeated-measures ANOVA for Mtot showed no significant main effect for either intensity level [F(2.043,42.894) = 0.456, p = 0.641, ηp2 = 0.021] or time [F(1,21) = 1.213, p = 0.283, ηp2 = 0.055]. However, there was a significant intensity*time interaction present [F(3,63) = 6.379, p = 0.001, ηp2 = 0.233]. As can be seen in Table 2, the Mtot response was different when comparing the intensity conditions, and there was a significant increase found only in the RPE17 condition. Similarly, for Mdi, there was not a significant main effect for either intensity level [F(3,63) = 1.566, p = 0.206, ηp2 = 0.069] or time [F(1,21) = 1.851, p = 0.188, ηp2 = 0.081]. The interaction between the two independent variables (intensity*time) was significant [F(3,63) = 3.224, p = 0.028, ηp2 = 0.133]. The only significant simple effect, however, was found for the SPV test (Table 3).

Table 1 Results of the self-paced VO2max (SPV) test
Full size table
Table 2 CFF testing results: Mtot
Full size table
Table 3 CFF testing results: Mdi
Full size table

Discussion

The intention of the present study was to investigate the effect of two different types of physical exercise on cortical neural arousal. We hypothesized that the change in CFF variables would be different when comparing short, fatiguing exercise to longer, steady-state exercise. The results supported this hypothesis, and different responses were in fact elicited between the exercise types. The short, fatiguing exercise produced no change in Mtot but a significantly different Mdi value; the longer, steady-state exercise had contrasting results, seemingly once the threshold for intensity (RPE17) was met. These results suggest that in healthy recreational runners, fatiguing exercise and steady-state exercise at a constant RPE13 and 15 does not increase neural arousal or the rate of information processing. To achieve a higher rate of processing, vigorous intensity steady-state exercise (RPE17) may be necessary.

Increases in neural arousal observed at RPE17 may be attributed to many physiological mechanisms. Potential factors include relative changes in circulating catecholamines (McMorris and Hale 2015) as well as increases in brain perfusion (Ogoh and Ainslie 2009; Smale et al. 2017). Exercise intensity has been found to modulate the sympathoadrenal system (SAS) and hypothalamic–pituitary–adrenal axis (HPA) (McMorris et al. 2009) which could increase cortical neural arousal. Given that epinephrine and norepinephrine produce many stimulatory effects on substrate metabolism during glycolysis, lypolisis, and amino acid degradation (Zouhal et al. 2008), circulating catacholamines are often cited as a mechanism for producing auxiliary increases in cognitive function. Moreover, increases in cerebral blood flow (CBF) likely mediate oxygen perfusion to cortical areas involved with information processing. Based on the current findings, steady-state exercise at a vigorous intensity of RPE17 may have stimulated these hormonal and vascular changes and lead to the observable effects in Mtot.

Failure to identify any change in neural arousal after steady-state exercise at RPE13 and 15 may be the result of an insufficient stress stimulus to the neuroendocrine system. As previously mentioned, increases in circulating catacholamines have been purported to enhance cognitive function. Furthermore, increasing evidence suggests that a threshold exists as to when the neurocognitive benefits of exercise will occur. For example, McMorris and colleagues (2009) found a linear increase in circulating catacholamines from rest to 50% maximum aerobic power (MAP), and further increases from 50 to 80% MAP; however, no improvements in reaction time (RT) were noted at 50% MAP when compared to rest. Prior studies that have shown no change or a decrease in Mtot may have implemented intensity levels that were too low to elicit a positive improvement. For example, Loy and O’Connor (2016) used 30 min of cycling at only 40% VO2max, Suvi et al. (2016) used walking at 60% VO2max, and Grego et al. (2005) had subjects cycle at 60% VO2max. In the current study, the biggest and only significant increase in Mtot was found at RPE17, which corresponds to approximately ≥ 85% VO2max (Thompson et al. 2009). The results from this study, and the previously mentioned studies, suggest that intensity is an important determinant of cortical neural arousal increases.

Though most studies have shown a decrease in cognitive function when exercise intensities approach maximal effort (LaManca et al. 1998; Tomporowski 2003), one study has produced conflicting results. Davranche and Pichon (2005) had subjects perform a cycle-based VO2max test, and assessed CFF before and after exercise. They hypothesized a transitory CNS fatigue would be induced, leading to a decrease in sensory sensitivity, thus a lower CFF threshold value. However, their results showed an increase in arousal, as shown by a significantly higher Mtot value after exercise. However, Godefroy et al. (2002) and the current study used a different modality, running, and did not find any differences in Mtot pre/post-exercise. This suggests that modality may affect neural arousal in some fashion. There are limited studies to which we can compare results. Therefore, more research in the area of maximal exercise and cognitive function is warranted, especially with the popularity of high intensity interval training (HIIT).

Exercise necessitates the integration of nearly every system in the body. The demands of physical movement via neuromuscular coordination and activation of the associated motor cortices, basal ganglia, cerebellum, and sub-nuclei may encroach on available cognitive resources committed to higher order brain processes (Dietrich 2003). As a result, cognitive tasks requiring increased mental effort such as working memory, ability to shift cognitive sets, and sustained attention may be affected (Dietrich and Sparling 2004). The transient hypofrontality theory (Dietrich 2003) bolsters this claim by suggesting that brain regions anterior to the central sulcus (prefrontal cortices), that are specifically involved in higher cognitive processes, are downregulated to support lower level neural circuitry involved with movement. As a result, mental tasks requiring substantial focus and effort are negatively affected. This may be a plausible explanation as to why Mtot showed a slight decrease after the short, fatiguing exercise. Further support for this line of thought is confirmed in a study conducted by Covassin et al. (2007), who showed decreased neurocognitive function after VO2max testing. Ultimately, it seems that brain regions involved with higher order processes, such as the mechanical and musculoskeletal demands of a VO2max test, display a tradeoff contingency. As exercise intensity, duration, and the amount of musculature involved in the activity increase, brain regions primarily responsible for tasks not related to movement decrease. Within the context of the current study, visual acuity sensitive to changes in flickering stimuli was affected immediately following short, but fatiguing exercise. The significant change in the response criterion (Mdi) after the SPV may be explained by the role that the anterior cingulate cortex plays in decision-making and response inhibition (McMorris et al. 2009; Wittmann et al. 2016). The large Mdi values before the SPV could indicate an anticipatory component to the maximal exercise test that the subjects would soon be completing. McArdle et al. (1967) showed that prior to a running event, increases in heart rate at rest corresponded to expected intensity level of the event; the highest pre-exercise heart rates were found before the shorter, and thus more intense, events. It is believed that this is due to a decrease in vagal tone, an increase in sympathetic activity, and the motor cortex priming the body for intense exercise. Our results are in line with Godefroy and colleagues (2002), who showed a significant change in Mdi but not in Mtot after a VO2max test. These results suggest that a short, fatiguing bout of exercise does not significantly alter neural arousal, but can have a profound effect on decision-making skills.

As previously stated, there are many theories that explain how higher order cognitive function is diminished following exercise. Some attribute the dispersal of resources to neurotransmitter modulation (Gomez-Pinilla and Hillman 2013; McMorris and Hale 2015; Yanagisawa et al. 2010), while others take into account CBF changes (Guiney and Machado 2013). Additional factors to consider are the mode of exercise (Lambourne and Tomporowski 2010), initial fitness level (Chang et al. 2012), CFF starting frequency, and the timing at which cognitive assessments take place. When considering mode of exercise, results form a recent meta-analysis (Lambourne and Tomporowski 2010) showed larger effect sizes associated with cycle ergometry when compared to the treadmill. The authors speculated that differences in cortical activation specific to the musculature involved in each task could have contributed to these differences. Fitness level has also been suggested as a moderator of cognitive performance during testing (Chang et al. 2012), though inconclusive findings have made a consensus on this point ambiguous at best. There have also been conflicting results when investigating neural arousal before and after a bout of exercise. Inconsistency can be attributed to the range of methods used when assessing cortical arousal, such as the starting frequency for the trials. Though we used 12 and 50 Hz, there have been many others starting frequencies used. Zero and 100 Hz (Davranche and Pichon 2005; Lambourne et al. 2010), 0 and 44 Hz (Grego et al. 2005), 20 and 44 Hz (Godefroy et al. 2002) and alternating starting frequencies (Presland et al. 2005) have all been used. Some studies do not state the starting frequencies at all (Loy and O’Connor 2016; Suvi et al. 2016). Disparate findings among researchers have also led many to consider the time at which cognitive function is evaluated (i.e., prior to, during and after exercise). Though some argue that the timing is essential and that assessments should be implemented during and post-test, a recent meta-analysis by McMorris and Hale (McMorris and Hale 2012) found no differences between effect sizes for speed of processing when comparing post-exercise and during-exercise measurements.

Conclusion

Our results suggest that short, fatiguing and longer, steady-state exercise affect cortical neural arousal differently. Increases in arousal, and the rate of information processing, are more likely to come from longer, steady-state exercise at a vigorous intensity than from short, fatiguing exercise at a maximal intensity. Future research should delve deeper into the relationship between neural arousal and exhaustive exercise when subjects are not allowed to self-pace. Furthermore, auxiliary testing techniques including electroencephalography (EEG) and functional Magnetic Resonance Imagery (fMRI) could be used in conjunction with CFF to provide a more thorough description of relationship between exercise and neural arousal.