Time-of-Day Effects on Short-Duration Maximal Exercise Performance

Time-of-day dependent fluctuations in exercise performance have been documented across different sports and seem to affect both endurance and resistance modes of exercise. Most of the studies published to date have shown that the performance in short-duration maximal exercises (i.e. less than 1 min - e.g. sprints, jumps, isometric contractions) exhibits diurnal fluctuations, peaking between 16:00 and 20:00 h. However, the time-of-day effects on short duration exercise performance may be minimized by the following factors: (1) short exposures to moderately warm and humid environments; (2) active warm-up protocols; (3) intermittent fasting conditions; (4) warming-up while listening to music; or (5) prolonged periods of training at a specific time of day. This suggests that short-duration maximal exercise performance throughout the day is controlled not only by body temperature, hormone levels, motivation and mood state but also by a versatile circadian system within skeletal muscle. The time of day at which short-duration maximal exercise is conducted represents an important variable for training prescription. However, the literature available to date lacks a specific review on this subject. Therefore, the present review aims to (1) elucidate time-of-day specific effects on short-duration maximal exercise performance and (2) discuss strategies to promote better performance in short-duration maximal exercises at different times of the day.

Circadian rhythms are responsible for temporal regulation of numerous physiological phenomena in the human body. At a molecular level, circadian rhythms are defined as a function of clock gene expression levels over a 24-hour period 1 . Clock gene expression levels throughout the day set up the pace of sleep-wake cycles, hunger, hormone production, body temperature, as well as other physiological functions 2 . In mammals, biological rhythms are driven by a central pacemaker located in the suprachiasmatic nuclei (SCN) of the hypothalamus. It consists of approximately 20 000 neurons that exhibit independent rhythms of firing rate and gene expression 3 . The core clock system of the SCN works as a self-sustained transcriptional/translational feedback loop, involving a set of four integral proteins that act as activators or repressors within the system 4 . In addition, there is a group of kinases and phosphatases that regulate their localization and stability 5 . Apart from the core clock system in the SCN, circadian clocks and clock-controlled output genes (CCGs) are also present in peripheral tissues such as the liver, heart, kidney, pancreas, adipose tissue and skeletal muscle [6][7][8][9][10][11][12] . In this sense, the main synchronizer of the internal clock is solar light 13 , but it is known now that other non-photic stimuli such as feeding, social contact or physical exercise do also modulate the transcriptional activity of clock genes [14][15][16][17] . Likewise, internal clocks also play important roles on human behavior, accounting for variations in resting levels of neuromuscular, sensory-motor and cognitive performance throughout the day [18][19][20][21] . Recently, high-throughput transcriptomic and metabolomic analyses in mice have demonstrated that the time of day is a crucial factor to amplify the effect of exercise on systemic energy homeostasis and metabolic pathways within skeletal muscle 22 . In addition, exercise capacity exhibit diurnal fluctuations in mice and humans between the early and late part of their active phase 23 , which suggests that the time of day is a major modifier of exercise capacity.
Time-of-day dependent fluctuations in exercise performance have been documented across different sports and seem to affect both endurance and resistance modes of exercise [24][25][26][27][28][29][30] . Whereas differences in endurance

Results
A total of 66 articles were selected and then divided into dynamic and/or isometric/isokinetic short-duration maximal exercises. Forty-four articles assessed time-of-day effects on dynamic short-duration maximal exercise performance, including swimming, tennis, jumping, cycling, sprinting and resistance exercises (Table 1). On the other hand, 32 articles assessed time-of-day effects on isometric and isokinetic exercise performance, including electrically induced contractions, reflex contractions, maximal and submaximal voluntary isometric contractions, isokinetic leg contractions and magnetic stimulation of the cortex (Table 2). Most articles compared time-of-day effects on short-duration maximal exercise performance under acute exercise conditions; however, 11 articles assessed time-of-day effects on short-duration maximal exercise performance after a period of training (from 5 to 10 weeks) (Tables 1 and 2).
Neuromuscular differences in dynamic short-duration maximal exercise performance throughout the day has been explored using electromyographic (EMG) activity recordings. Castaingts et al. 55 analyzed variations in force and EMG activity of skeletal muscles throughout the day as well as the ratio between these parameters (i.e. force/EMG activity) during a natural movement (i.e. drop jump). Such relationship between the force and EMG activity is called neuromuscular efficiency (NME). In this study, they observed a higher jump height and NME in the evening than in the morning, which indicates that the process of storage-release of potential energy in muscle elastic elements is improved in the evening, without a parallel increase in motor unit activation 55 . A further analysis of time-of-day fluctuations in power and EMG activity of vastus lateralis, rectus femoris, vastus medialis and biceps femoris muscles recorded in a repeated sprint exercise protocol (5 × [6 s of maximal cycling sprint + 30 s of rest]) showed that total work, percentage of peak power decrement and peak power were higher in the evening than in the morning, although it was not accompanied by a time-of-day effect on EMG activity levels 44 . Similarly, Chtourou et al. 24 recorded EMG activity changes during a Wingate anaerobic test and showed that power output and NME were higher in the evening during the first phase of the test (i.e. first 20 s), where peripheral mechanisms of muscle contraction have the main role in exercise performance. Yet, they were independent of the time of day during the second phase of the test (last 10 s), where central mechanisms of muscle contraction have a higher role in exercise performance. Thus, since most studies have shown that changes in muscle function throughout the day are not accompanied by changes in EMG activity levels, it has been suggested that adaptations at the muscle fibre level rather than changes of the neural drive, motor unit properties, and/or muscle membrane properties are more likely to cause time-of-day dependent variations in dynamic short-duration maximal exercise performance 66 .
Other studies have also suggested the existence of peripheral mechanisms which might explain diurnal oscillations in dynamic short-duration maximal exercise performance. For instance, Hammouda et al. 58 observed that the neuromuscular performance during a repeated sprint ability test (5 × [6 s of maximal cycling sprint + 24 s of rest]) was higher in the evening and it was accompanied not only by higher levels of biological markers of muscle injury but also a lower antioxidant status at this time of the day. Furthermore, Racinais et al. 61 assessed time-of-day differences on repeated sprint exercise performance and found a better performance and higher blood lactate concentrations in the evening than in the morning. Although the physiological basis of these time-of-day dependent oscillations has not been elucidated, the evidence suggests the existence of a muscle specific mechanism that accounts for time-of-day dependent fluctuations in neuromuscular performance, whose activity is at least partially independent of the central nervous system.
The time of day influences isometric and isokinetic exercise performance in both lower and upper extremities. Regarding lower extremities, Guette, Gondin and Martin 26 found a significant time-of-day effect on maximal voluntary muscle contraction (MVC) peak torque of the quadriceps muscles on the dominant and non-dominant leg, with the highest values occurring at 18:00 h. Regarding upper extremities, Gauthier et al. 77 observed a time-of-day dependent rhythm in elbow flexor torque, whose acrophase was reached at around 18:00 h. This was accompanied by a time-of-day dependent rhythm in biceps muscle EMG activity. Thus, it is clear that upper and lower limb muscle contractility in isometric and isokinetic actions is affected by the time of day. In this sense, most of the studies have shown that when the performance is assessed at a peripheral level by means of peak power, mean power or total work, then a typical peak in performance is found in the evening 29 www.nature.com/scientificreports www.nature.com/scientificreports/ However, the ability to generate force in isometric and isokinetic exercises depends not only on peripheral but also on central mechanisms of muscle contraction. Central mechanisms include central nervous system command, alertness and motivation, being them all normally assessed by electromyographic activity recording of skeletal muscles 68 . When muscle performance is assessed at central level, higher morning values 77,79 , higher evening values 80 or no differences throughout the day 26,29,66,82 are observed. Such differences in EMG activity throughout the day observed among different studies are difficult to be explained. However, many researchers suggest that EMG activity recordings can be affected by the muscle groups examined. Strikingly, time-of-day dependent fluctuations in maximal isometric and isokinetic performance seem to be different between fast and slow muscles of the lower limbs. In this sense, a decrease in MVC peak torque of triceps surae muscles and soleus muscle EMG activity was observed in the evening in comparison to the morning. According to the authors, this is probably due to a higher fatigue state of the slow motor units as the day progresses 39,79 . It is worth noting that methodological factors associated with EMG activity recordings might also account for differences in central mechanisms of muscle contraction throughout the day.
Finally, some studies have suggested that diurnal fluctuations in short-duration maximal exercise performance may be partially controlled by the individual circadian typology/chronotype. Thus, inter-individual differences related to the chronotype might generate opposite responses in central and peripheral mechanisms of muscle contraction throughout the day. Chronotype is an individual's characteristic pattern that reflects preferences towards morningness or eveningness, and it is usually evaluated using self-assessment questionnaires. The effect of the chronotype on isometric and dynamic exercise performance has not been extensively studied, but some studies have shown the importance of the individual typology on isometric exercise performance. For instance, Kuusmaa et al. 68 showed that morning-type (M-type) individuals exhibited lower force values in the evening whereas evening-type (E-type) individuals exhibited lower force values in the morning, despite the fact that EMG activities where not different throughout the day in either group. Tamm et al. 75 reported that E-type individuals showed parallel increases in cortical and spinal excitability over the day, which were reflected on an increased triceps surae EMG activity and MVC torque at 21:00 h. However, M-type individuals exhibited the highest cortical excitability at 09:00 h but the highest spinal excitability at 21:00 h, without significant differences in triceps surae EMG activity or torque produced during MVCs over the day. Such discrepancies have led some scientists to the conclusion that EMG activity and MVC torque could increase, remain constant or even decrease throughout the day, thus reflecting an orchestrated response between peripheral and central mechanisms in the control of skeletal muscle contractile properties 55 .
Effects of temperature. Since short-duration maximal exercise performance fluctuates in parallel with body temperature 56,62 , it has therefore been suggested that body temperature affects the contractile properties, viscosity and conduction velocity of action potentials in skeletal muscle as well as the extensibility of connective tissue 91 . In fact, Racinais and Oksa 92 have found a positive and linear relationship between performance and muscle temperature, where performance improves from 2 to 5% with a 1 °C increase in muscle temperature. Hence, several research groups have been devoted to studying the effects of getting exposed to warm or cold environments on short-duration maximal exercise performance at different times of the day. In this sense, it was observed that a 60-min exposure to a moderately warm and humid environment (i.e. 28.1-29.5 °C, 62.6-74% relative humidity) cycling sprint + 5 min of rest], 10 i.e. no medication within the last14 days, a non-smoker, regular sleep pattern with sleep duration ranging from 6 to 9 h per night and regular physical activity not more than once a week, 11 i.e. 12 min of pedaling at 50% of V˙O2max interspersed with three brief accelerations of 5 s, 12 i.e. 3 min of pedaling at 70 rpm at 50% of V˙O2max, 13. i.e. 60 min of resting in a room at 29 °C, 70% relative humidity, 14 i.e. 60 min of resting in a room at 28.1 °C, 62.6% relative humidity, 15 i.e. 60 min of resting in a room at 24 °C, 70% relative humidity, 16 i.e. 30 min exposure to a cold bath at 16 °C, 17 i.e. 30 min exposure to a bath at 38 °C. RPE rating of perceived exertion, P max Maximal Power, P peak Peak Power, P mean Mean Power, F max Maximal Force, V max Maximal Velocity, CMJ countermovement Vertical jump, SJ Squat Jump, MCV Maximal Voluntary Isometric Contraction, NME Neuromuscular efficiency, EMG Electromyographic activity, 1 i.e. 5 ×[6 s of maximal cycling sprint + 24 s of rest], 10 i.e. no medication within the last14 days, a non-smoker, regular sleep pattern with sleep duration ranging from 6 to 9 h per night and regular physical activity not more than once a week, 18 i.e. a minimum of 6 months resistance training history 19 i.e. 20 min general warm-up on a stationary bike at 150-200 W prior to completion of the control warm-up, resulting in a whole-body temperature increase of 0.3 ± 0.2 °C, * estimated acrophase by cosinor analysis. www.nature.com/scientificreports www.nature.com/scientificreports/ blunts the diurnal variation in muscle function that is observed in neutral climate conditions (i.e. 20-20.5 °C, 57-67% relative humidity) 56,62,71 . Conversely, immersion in cold water (i.e. 16-16.5 °C) before performance testing in the evening has shown to reduce the force and muscle power towards morning values 60,72 . This is achieved through a "passive effect" of thermal conditions on muscle temperature 92 . In addition to the passive effect of climate conditions on body temperature, time-of-day dependent variations in muscle performance might also diminish or even disappear when extended warm-up protocols are used (i.e. 20 min cycling on a stationary bike at 150-200 W prior to the completion of the regular warm-up) 57 . Diurnal fluctuations in performance also disappear under active (i.e. 12 min of pedaling at 50% of VO 2max interspersed with three brief accelerations of 5 s) but not passive (i.e. 3 min of pedaling at 70 rpm at 50% of VO 2max ) warm-up conditions 59 . The same can be stated when longer active warm-ups (i.e. 15 min of pedaling at 50% of maximal power output) are compared with shorter ones (i.e. 5 min of pedaling at 50% of maximal power output) 42 . The effectiveness of these protocols to increase performance in the morning towards evening values is due to an "active effect" on muscle temperature 92 . Thus, increases in muscle temperature might help athletes to minimize diurnal dependent variations in short-duration maximal exercise performance. This can be achieved by means of short exposures to warm and humid environments as well as by the inclusion of active warm-up protocols (e.g. 12-15 min pedaling at minimum intensities of 50% VO2max and interspersed with sprint exercises) before exercising.
Finally, although there is strong evidence to support that time-of-day dependent fluctuations in short-duration maximal exercise are dependent on body temperature, it is clear that there are additional factors which might affect such fluctuations. In this sense, it should be noted that in moderately warm and humid climate conditions better performances are achieved in the morning, when body temperature is at its lowest. However, this is not observed in the evening 56,62,71 . In this regard, it seems that the effects of the time of day on body temperature and those of warm and humid climate conditions do not interact to potentiate exercise performance in the evening. According to some authors, this is due to the similar effects of both variables on neuromuscular efficiency 59 , although there is no data to support this hypothesis. On the other hand, it has been observed that an increase in rectal morning temperature towards evening values does not increase muscle strength levels in the morning towards levels found in the evening 76,87,89 . This might suggest that a specific mechanism within the muscle is orchestrating the effects of time of day on short-duration maximal exercise performance. It is also worth mentioning that when passive exposure to hot environments increases the central temperature towards values close to 39 °C, there is a reduction in the neural drive and the maximal voluntary activation of muscles. This will negatively affect power production and short-duration maximal exercise performance 92 .
Effects of music. In addition to peripheral factors, it has been suggested that central factors related to alertness, motivation and mood might affect short-duration maximal exercise performance at different times of the day 13,28,33,86,93,94 . It has been well established that an auditory stimulus, such as music, is effective in increasing levels of arousal, reducing levels of perceived exertion and facilitating a better motor coordination during short-duration maximal exercises [94][95][96][97][98] . In this regard, Chtourou et al. 13 observed higher performances in the Wingate anaerobic test when it was performed after warm-up with music [10-min warm-up while listening high tempo music (>120-140 bpm) through headphones] in comparison to warm-up without music conditions. Furthermore, listening to music during the warm-up period blunted time-of-day dependent variations in muscle power output. This effect was achieved due to a greater increase in performance in the morning than in the evening. Interestingly, a recent study carried out by the same research group showed that this beneficial effect of music on short-duration maximal exercise performance can be achieved when the 10-min warm-up is performed either under neutral (i.e. non-selected) or self-selected motivational music conditions 86 . In the same study, a greater effect of music on short-duration maximal exercise performance in the morning than in the evening was reported. According to the authors, the ergogenic effect of music could be related to better individual perceptions of self-esteem and sense of confidence in the morning, where the level of motivation seems to be lower 45% relative humidity, * estimated acrophase by cosinor analysis. EMG electromyographic activity, MCV maximal voluntary contraction, NME Neuromuscular efficiency (mechanical performances/EMG recordings), MVCLP maximal bilateral isometric leg press, EMGLP myoelectric activity during maximal bilateral isometric leg press, MVCKE maximal unilateral isometric knee extension, MVCVA maximal voluntary activation level unilateral isometric knee extension, EMGVA myoelectric activity during maximal voluntary activation level unilateral isometric knee extension, VA% voluntary activation percentage of the quadriceps muscles, * estimated acrophase by cosinor analysis. EMG electromyographic activity, MCV maximal voluntary contraction, NME Neuromuscular efficiency (mechanical performances/EMG recordings), N/A Not available, 6 i.e. knee extension at 2.09 rad.s(-1) through a 90° range of motion, 7 i.e. knee flexion at 240°/s, 60°/s, 0°/s, −60°/s, 8 i.e. knee flexion and extension at 90°/s, 80°/s, 70°/s, 50°/s and 30°/s, 9 i.e. knee flexion and extension at 1.05 and 4.19 rad.s(-1) through a 90° range of motion, 10 i.e. no medication within the last14 days, a non-smoker, regular sleep pattern with sleep duration ranging from 6 to 9 h per night and regular physical activity not more than once a week, 111 i.e. immersion in cold water (16.5 °C) before testing to lower rectal temperature to morning values, 12 i.e. 60 min of resting in a room at 29.5 °C, 74% relative humidity, 13 i.e. 60 min of resting in a room at 20.5 °C, 67% relative humidity. EMG electromyographic activity, MCV maximal voluntary contraction, 10 i.e. no medication within the last14 days, a non-smoker, regular sleep pattern with sleep duration ranging from 6 to 9 h per night and regular physical activity not more than once a week. MVF maximal voluntary force, MRFD maximal rate of force development, 14 i.e. knee flexion and extension at 1.05 and 3.14 rad.s(-1) through a 90° range of motion. (2020) 10:9485 | https://doi.org/10.1038/s41598-020-66342-w www.nature.com/scientificreports www.nature.com/scientificreports/ among individuals. Thus, listening to music during the warm-up period could be an effective strategy to improve short-duration maximal exercise performance, especially in the morning hours.
Effects of testosterone and cortisol. Short-duration maximal exercises produce an acute increase in both serum testosterone and cortisol levels, affecting the anabolic and catabolic status in skeletal muscle 99 . In addition, testosterone and cortisol are under circadian regulation, both peaking in the morning hours 100 . This has raised the question of whether the exercise-induced hormonal response can be influenced by the time of day. In this regard, Bird and Tarpenning 101 found that a single session of heavy resistance exercise produced a lower cortisol response when performed in the evening compared with the morning, without affecting the diurnal rhythmicity of testosterone. This suppression in the cortisol response after resistance training in the evening hours was also observed by Burley et al. 102 . The authors proposed that this reduction in the testosterone/cortisol ratio leads to a reduced catabolic environment which could favor muscle adaptations to resistance exercise in the evening hours. Interestingly, Sedliak et al. 73 observed that a 10-week training period at a specific time of day gradually reduced morning but not evening cortisol levels. The authors attributed these changes to a lower anticipatory stress and a better accommodation to early waking rather than any chronic change in cortisol diurnal rhythmicity. However, this period of training at a specific time of day was enough to blunt the time-of-day effects on strength performance in these individuals. Moreover, the absolute increase in maximum strength at the end of the training period was similar in all individuals regardless the time of day at which training was conducted.
It is worth mentioning that, although some studies have observed a relationship between hormone levels, short-duration maximal exercise performance and the time of day, there are other ones that have not so 63,64,74 . For instance, Kūūsma et al. 64 observed no changes in the typical diurnal variations in resting serum testosterone and cortisol concentrations after a 24-week period of training at a specific time of day. According to the authors, the effects of a temporary phase shift caused after a prolonged period of training might explain these results. Furthermore, Sedliak et al. 74 observed that resting cortisol levels declined independently of the time of day after 11 weeks of training at a specific time of day in untrained young males. However, they did not find changes in resting testosterone levels after the training period in such individuals. Likewise, no significant changes were observed in resting testosterone levels in a group of healthy older women after 12 weeks of training at a specific time of day 63 . In summary, whereas resistance exercise training periods at a specific time of day do not seem to affect the circadian pattern of testosterone secretion, their effect on cortisol secretion diurnal pattern is more controversial. Therefore, further research is required to elucidate the role of the time of day on the exercise-induced hormonal response.
Effects of circadian systems. Recently, a large-scale transcriptomic analysis has revealed the existence of rhythmic and CLOCK-driven pathways in human skeletal muscle that affect ~8% of muscle genes 103 . This analysis showed that the transcription of human skeletal muscle clock genes was distributed into two phases of transcript accumulation at 04:00 and 16:00 h. The afternoon peak was enriched in genes related to muscle contraction and mitochondrial activity whereas the early morning peak was enriched in genes associated with inflammation and the immune response 103 . In addition, the use of genomic techniques has led to the identification of metabolic pathways regulated by clock genes within skeletal muscle, including those involved in glucose uptake, lipid metabolism and myokine secretion [104][105][106] . Dyar et al. 104 observed that skeletal muscles of mice with a specific ablation of the core clock gene Bmal1 present an impaired insulin-stimulated glucose uptake due to (1) reduced protein levels of the insulin-dependent glucose transporter GLUT4 and the Rab-GTPase TBC1D1 involved in GLUT4 translocation to the plasma membrane and (2) a decrease in the activity of the glycolytic enzyme PDH. Furthermore, a high-resolution microarray analysis in a muscle-specific Bmal1 knockout mice model revealed a temporal separation of genes involved in carbohydrates and fatty acids use and storage over a period or circadian time of 24 h (CT 1 to 24) 107 . In this sense, circadian expression of genes involved in carbohydrate catabolism such as hexokinase-2 (Hk2; which catalyzes the first step of glycolysis) and pyruvate dehydrogenase phosphatase (Pdp1; which activates PDH) peak in the late inactive and early active phases of the day (CT 10-12). Moreover, the pyruvate dehydrogenase kinase gene (Pdk4; which inhibits PDH) peaks in the middle of the inactive phase of the day (CT 6). This circadian pattern of gene expression might promote an increase in the glycolytic flux during the active phase of the day (CT 12-24) 107 . On the other hand, genes involved in fatty acid uptake and β-oxidation such as acyl-carnitine translocase (Slc25a20; which transfers fatty acids into the mitochondrial matrix) and malonyl-CoA decarboxylase (Mlycd; which promotes β-oxidation by reducing malonyl-CoA levels) reach their peak expression in the middle of the inactive phase of the day (CT 7.5). Moreover, the gene that encodes for the nuclear PPARG co-activator 1 beta (Ppargc1b) which stimulates the activity of several transcription factors involved in mitochondrial biogenesis, fatty acid uptake and β-oxidation, also peaks in the middle of the inactive phase of the day (CT 7) 107 . Altogether, these results could suggest that there is a shift in the use of substrates from lipids to carbohydrates. This metabolic shift, which depends on the circadian gene expression in skeletal muscle, goes from the inactive (CT 1-12) to the active (CT 12-24) phase of the day. This circadian pattern of metabolic gene expression in skeletal muscle could explain the greater performance in short-duration maximal exercises observed in the evening hours (i.e. between 16:00 and 20:00 h). In this sense, a greater activation of the glycolytic pathway in the active phase of the day would promote greater performances in anaerobic exercises in the evening hours.
In spite of the typical temporary peak observed in the evening hours, many studies have demonstrated that regular training at a specific time of the day is able to blunt the diurnal fluctuations in maximal exercise performance 30,36,45,54,[63][64][65][66]69,73,74,78 . The mechanisms responsible for the adaptations of training at a specific time of day in short-duration maximal exercise performance remain unknown. However, Sedliak et al. 66 observed that the EMG activity during a maximal voluntary contraction of the knee extensors did not show any time-of-day specific adaptation after a 10-week training period at a specific time of day. This led the authors to suggest that peripheral (2020) 10:9485 | https://doi.org/10.1038/s41598-020-66342-w www.nature.com/scientificreports www.nature.com/scientificreports/ rather than neural adaptations are the main source of temporal specificity in strength training. This could be caused by (1) an attenuation in training adaptations throughout the 10-week training period or (2) individual chronotype-related differences in responsiveness to training at a specific time of day. Likewise, intermittent fasting conditions, such as those imposed during the Ramadan period (i.e. 15-16 h starvation/day; from ≈ 04:00 h until ≈ 19:00 h) have shown to exert the same effect, minimizing the time-of-day dependent variations in exercise performance 50,85 . Chtourou et al. 50 have suggested that Ramadan might act directly on the circadian rhythm of anaerobic performance by means of (1) inducing a phase advance or delay in the rhythm or (2) reducing the amplitude of the rhythm of the anaerobic power development. Despite the fact that none of these hypotheses have been tested to date, some of them have raised the question as whether circadian systems could account for time-of-day specific training adaptations.
There is evidence that scheduled exercise is able to cause phase shifts in the circadian system within skeletal muscle in mice 15,16 . Using a PER2::LUC circadian reporter mouse model, Wolff and Esser 15 showed that four weeks of either voluntary or involuntary exercise for 2 hours/day were able to shift the phase of locomotor and molecular rhythms in skeletal muscle towards an earlier time of day. Furthermore, Edgar and Dement 16 used two different wheel restriction schedules to determine whether exercise duration was an important determinant of feedback activity to the mouse circadian system. In this study, they found that more than 75% of total wheel activity occurred in the initial 2 hours over a 6-h or 12-h period of wheel availability. This led the authors to suggest that exercise intensity might be a primary determinant of the exercise-dependent phase shift strength. Thus, the ability of exercise to provide temporal feedback to the circadian system could explain why time-of-day effects on short-maximal exercise performance are blunted after a training period at a specific time of day. In humans, Zambon et al. 17 found that the expression of circadian clock genes was affected in quadriceps muscles at 6 and 18 h after having performed a single session of one-leg resistance exercise. In this study, resistance exercise shifted the expression patterns of diurnal-regulated genes through two mechanisms: (1) by upregulating genes that are normally repressed in the morning; and (2) by downregulating genes that are normally activated in the morning. Among them, it was observed an upregulation of the gene that encodes for a regulatory subunit of protein phosphatase-1 (Ppp1r5), which is normally repressed in the morning. Protein phosphatase-1 is responsible for activating glycogen synthase while also inhibiting glycogen breakdown. Moreover, it was observed a downregulation of the gene that encodes for the mitochondrial uncoupling protein 3 (Ucp3), which is upregulated in the morning. Uncoupling protein 3 stimulates oxidative phosphorylation by creating proton leaks across the inner mitochondrial membrane. These data show that exercise represents an important circadian time cue and changes the phase of the molecular clock, specifically in peripheral tissues. Altogether, these results suggest that skeletal muscle molecular clocks might account for the adaptations of training at a specific time of day, suggesting a link between molecular clocks in skeletal muscle and exercise performance throughout the day.

Practical applications.
Based on the studies summarized in this manuscript, it is clear that the time of day at which short-duration maximal exercise is conducted is an important variable for training prescription. However, to date, there are no guidelines to help athletes or coaches to achieve optimal performances or avoid deleterious effects of time of day on short-duration maximal exercise performance. Accordingly, some general remarks in this respect include: (1) under neutral climate conditions (i.e. 20-20.5 °C, 57-67% relative humidity), better short-duration maximal exercise performances are achieved in the afternoon (i.e. between 16:00 and 20:00 h) compared with the morning; (2) time-of-day deleterious effects on short-duration maximal exercise performance in the morning may be minimized after: (a) a 10-min warm-up period while listening to neutral or self-selected high-tempo music (>120-140 bpm) through headphones; (b) a 60-min exposure to warm and humid climate conditions (i.e. 28.1-29.5 °C, 62.6-74% relative humidity); (c) active warm-up protocols (e.g. 12-15 min pedaling at minimum intensities of 50% VO2max and interspersed with 5-s sprint exercises); (d) 2-4 weeks of intermittent fasting conditions of 15-16 h starvation/day from ≈04:00 h till ≈19:00 h (e) a training period of at least 5 weeks performed in the morning.

Conclusion
In summary, this review shows that, under neutral climate conditions, short duration maximal exercise performance is affected by the time of day, peaking between 16:00 and 20:00 h. However, a similar performance may be achieved in the morning hours if exercise is conducted after: (1) short exposures to moderately warm and humid environments; (2) active warm-up protocols; (3) intermittent fasting conditions; (4) warming-up while listening to music; (5) prolonged periods of training at a specific time of day. This suggests that time-of-day dependent fluctuations in short-duration maximal exercise performance are controlled not only by body temperature, hormone levels, motivation or mood states but also by a versatile circadian system within skeletal muscle.

Data availability
The datasets generated during and/or analyzed during the current study are available in the MEDLINE and Google Scholar repositories, https://www.ncbi.nlm.nih.gov/pubmed/, https://scholar.google.com/, respectively.