KEY WORDS: Prepubertal and adolescent athletes, trainability, mechanisms, resistance training, anaerobic training, aerobic training, overtraining.

Can children be physically trained? Is training healthy for children? Can a child become overtrained? These are just some of the questions that exercise and sports scientists have debated over in the past; however conclusive answers to these issues remain elusive.

As Armstrong and Welsman, 2002 state, "children are not mini-adults" indicating that our understanding of the exercise physiology of an adult cannot just be scaled down and applied to children; Indeed the very process of growth and maturation complicates our understanding of the trainability of children as this process mimics the effects of training (Baxter-Jones et al., 1995; Naughton et al., 2000) and thus needs to be accounted for when determining the success or failure of an exercise training programme. Additionally with the majority of studies being cross-sectional in nature it is difficult to partial these aforementioned maturational effects on the training response. With these caveats understood the available literature suggests that children can show improvements in strength (Falk and Tenembaum, 1996; Payne et al., 1997; Tolfrey, 2007), aerobic fitness (Baquet et al., 2003; Baxter-Jones et al., 1993; Kobayashi et al., 1978), and anaerobic fitness (Rowland, 2005; Tolfrey, 2007) with exercise training.

In adults, training-induced adaptations to aerobic training have been extensively studied, and it is well documented that adults can improve their cardiorespiratory (aerobic) fitness levels if they are given the appropriate training stimulus. The focus of child-based studies has largely centered upon improving fitness levels for either athletic performance or for its relationship with health outcomes. Typically the training programme design used with children mirrors that used for adults (Rowland, 1985), whether this is appropriate or not remains to be clarified, but evidence indicates that higher intensities may have to be used to initiate a response in children (Baquet et al., 2003; Massicotte and Macnab, 1974). Unfortunately, despite the abundant research on aerobic training, the relationship between exercise training and adaptations in aerobic metabolism are still not well understood in prepubertal and early pubertal children (Baxter-Jones and Mundt, 2007; Tolfrey, 2007).

Early investigations have generally assumed that young children would have minimal or no response in aerobic fitness (peak VO₂) to endurance training, which was attributed to either a high inherent level of physical activity (American Academy of Paediatrics, 1976; Yoshida et al., 1980), or to some unexplained limitations in the biological responsive mechanisms related to maturation (Katch, 1983). This apparent "inferiority" of children to develop aerobic fitness was thought to be a reflection of the maturational differences between pre- and postpubertal children. The 'trigger hypothesis' (Katch, 1983) implied that there would be a time (trigger point) in a child's life - namely puberty - before which the effects of training would be negligible, or would not occur at all. The prepubertal hormonal environment of low GH and sex steroid concentrations were considered not suitable to bring about measurable improvements in physiological function and thus, enhanced aerobic fitness with training.

By how much can children improve aerobic fitness?
According to the literature it is believed that children show a reduced magnitude of improvements in aerobic fitness after endurance type training than that seen in adults. Young adults typically show an increase in peak VO₂ of around 15-20%, although a large intra- subject variation can be present (Bouchard et al., 1992), in children however this is reduced to <10%. The meta-analysis of Payne and Morrow (1997) analysed the data from 69 studies (of which 28 met the inclusion criteria) that were either a cross- sectional comparison between trained and untrained children or a pre-test/ post-test design. Greater differences between trained and untrained participants (Effect Size (ES) of 0.94 ± 1.00) were seen, but less than a 5% change in peak VO₂ (approximately 2mL·kg^-1·min^-1) were noted in the pre-post studies (ES 0.35 ± 0.82). More recently, Baquet et al., 2003 argued that improvements in peak VO₂ of around 5-6% are observed when the data are analysed independently by sex or pubertal status. These authors also stated that when only studies that reported significant changes were considered, peak VO₂ improvements rose to 8-10% (Baquet et al., 2003). In line with the two previous studies, Rowland, 2005 compiled a series of what he considered to be well-designed investigations in children and stated that the improvements in aerobic power were indeed quite small, with the greatest improvement being 10%, but the average change being 5.8%.

Training intensity
Is seems that children need to train at a higher exercise intensity to elicit increases in aerobic fitness than for adults. Heart rate has normally been used as a marker of training intensity, with recommended intensities between 60-90% of HR_max, or 60% of HR reserve (difference between resting and maximal HR) for adults. However, the same target heart rates do not seem to be effective in children as a sufficient exercise intensity is a necessary stimulus in order to see improvements in peak VO₂ (Rowland, 2005).

A study by Massicot and MacNab (1974) involving three separate groups of nine boys (11-13 years) set different training intensities for each group. Participants trained at heart rates of 130-140, 170-180, and 150-160 beats·min^-1, and were compared to a matched control group. After having been matched for baseline peak VO₂ and then randomly assigned to the four experimental groups, the boys undertook six weeks of aerobic training. The results showed that only the boys who exercised at the highest intensities (above 170 bts·min^-1 - at least 88% HR_max) showed an 11% increase in peak VO₂.

Indeed as long as the training intensity is high enough, improvements have been seen in both prepubertal and adolescent children., Baquet and colleagues (2002) looked at the effects of a 7-week (twice per week) high intensity intermittent training programme on peak VO₂ of prepubertal boys and girls. Peak VO₂ increased by 8.2% compared with a 1.9% reduction in the 20 maturity-matched control participants. The results demonstrated that prepubertal children can increase peak VO₂ with high intensity aerobic exercise. Also, the fact that children only exercised two times during the week comes to reinforce the importance of intensity as a major stimulus for peak VO₂ improvements.

Sex differences
At present, there are few studies, which have addressed sex differences in the trainability of aerobic fitness in children, with the majority of studies recruiting boys as participants. Single sex studies using girls however have reported that significant improvements in peak VO₂ are possible (McManus et al., 1997), akin to that seen in boys. The data from those studies that have directly compared the sexes suggest that boys and girls demonstrate similar responses in aerobic trainability (Baquet et al., 2003; Baxter-Jones et al., 1993), even though boys' peak VO₂ appears to be slightly higher than girls (Naughton et al., 1998; Obert et al., 2003).

Obert et al., 2003 had 10 girls and 9 boys participate in a 13-week endurance training programme (3 X 1 hour per week, intensity: 80%HR_max) and found that significant improvements in peak VO₂ were observed for both groups; boys increased peak VO₂ by 15% and girls by 8%. Similar findings were reported by Baquet et al., 2002; 9.5% increase in peak VO₂ for the prepubertal boys being not significantly different from the 7.2% increment observed in girls.

Can pre-pubertal children improve aerobic fitness?
To return to the hypothesis of Katch, 1983 discussed previously, it appears that prepubertal children are indeed able to show increased aerobic fitness levels with training (Baquet et al., 2002; Kobayashi et al., 1978; McManus et al., 1997; Obert et al., 2003; Tolfrey, 2007). Indeed, work by Danis et al. (2003) with sets of prepubertal and circumpubertal monozygotic twins - one acting as the control, the other undergoing endurance training for 6 months, 3 times per week, with the intensity set at 85-120% of their lactate threshold (LT) - saw improvements only in the prepubertal children and not the circumpubertal twins, clearly contradicting the postulations of Katch. Furthermore, one investigation looking at the effects of two types of training (cycling and sprint) during 8 weeks (3 times per week), on aerobic power of prepubertal girls showed significant increases in peak VO₂ (McManus et al., 1997).

The literature suggests that prepubertal and adolescent athletes are aerobically trainable, nevertheless, one should not disregard Katch's hypothesis since prepubertal children seem able to improve aerobic fitness to a lesser magnitude compared to adults.

Physiological mechanisms affecting aerobic fitness
What potential reasons have therefore been given to explain the reduced magnitude of the training response seen in children compared to adults?

Circulating testosterone
Adult data indicates that changes in the concentrations of hormones such as GH, IGF-I, testosterone, oestrogen are responsible for the anabolic effects seen with exercise training. It is not clear however how these hormones are affected by endurance training in children with much of the information being extrapolated from studies with adults and animals (Boisseau and Delamarche, 2000).

It appears however that testosterone concentration is markedly up regulated with training in children; sex-related increases in VO₂, muscle size and strength, and maximal arterio-venous oxygen (A-VO₂) difference, seem to be related with pubertal increases in testosterone. The time when testosterone starts increasing during puberty matches the age when aerobic trainability improves significantly (Rowland, 2005). Mero and colleagues (1990) measured resting serum testosterone, before, during and after a year of training in 11-12-year-old boys. The boys engaged in different activities like long distance running, sprint running, tennis and weight lifting. After the training period testosterone was almost 3 times higher in the training group compared to the untrained boys, and the mean testosterone level had approximately doubled after a year of training. It was concluded that the increases in physical performance and fitness were a direct result of the training induced enhanced testosterone concentration even in early puberty (Mero et al., 1990). Furthermore, there is evidence showing that testosterone levels can rise during acute bouts of exercise in adults, although not in prepubertal subjects (Fahey et al., 1979).

Cardiac changes
The cardiovascular characteristics of young adult endurance athletes have been well described. These athletes typically show a peak VO₂, which is approximately twice as large compared to the sedentary population, a reflection of a superior cardiac output (CO) resulting from a greater left ventricular size and consequent larger SV (Faria et al., 1989; Wolfe et al., 1986). Intervention studies confirm that SV and ventricular size are eminently trainable in adults, with SV increasing by 10-30% with exercise training (DeMaria et al., 1979; Stratton et al., 1994; Wilmore et al., 2001). An increase in peak VO₂ in adults is related to both central (increased maximal CO) and peripheral adaptations in A-VO₂ difference (Wilmore et al., 2001). In children, however, less information is available on this topic.

Cross-sectional studies of trained and untrained children suggest that cardiac size and function, thus ultimately SV, are greater in trained children. Nottin and colleagues (2002) found higher SV values in the trained cyclists compared to controls and stated that these increases were influenced by factors such as cardiac hypertrophy, augmented cardiac relaxation and possibly expanded blood volume. Furthermore, similar findings have been reported in trained versus untrained distance runners (Rowland et al., 1998).

Intervention studies also indicate that cardiac size and function are modifiable with training in children. Early work by Erickson and Koch (1973), using dye dilution technique, measured the cardiovascular responses of prepubertal boys (11-13-years-old) to 16 weeks of training. The authors observed decreases in resting HR, increases in heart size and peak VO₂ at the end of the programme; the latter related to the increase in SV rather than through an increased A-VO₂ difference (Eriksson and Koch, 1973). In agreement with earlier work, a recent study looking at cardiac adaptations after a 13-week endurance training programme in prepubertal boys and girls (10.5 ± 0.3 years), saw that peak VO₂ increased by 15% and 8% in the boys and girls respectively. The authors reported significant increments in SV (boys: +15%; girls: +11%) in both sexes, with this being the strongest explanatory variable for the increases in peak VO₂ (Obert et al., 2003).

Testosterone affects aerobic fitness by affecting changes in heart size (Janz et al., 2000). It has been shown that during puberty, the size of the left ventricle in males increases at a faster rate than it does in females (Hayward et al., 2001), which could work as an advantage for males during exercise by enabling increases in SV (stoke volume) and end-diastolic volume, and consequently CO (cardiac output).

With no difference in maximal A-VO₂ difference being reported in either cross-sectional comparisons or intervention studies in children, this indicates that central factors rather than peripheral usage may determine training induced adaptations in aerobic fitness. However it appears that the magnitude of change seen in cardiac factors is slightly lower than seen in adults and may contribute to the reduced aerobic trainability of children.

Aerobic enzymes
In adults, improvements in performance, lower submaximal exercise lactate production, and greater reliance on fat metabolism due to endurance training seem to be associated with an increase in mitochondrial numbers within the skeletal muscle (Holloszy and Coyle, 1984).

In one of the few studies in children to assess the changes in oxidative enzymes with aerobic training, Eriksson et al., 1973 looked at the effects 6 weeks of endurance training on muscle metabolic profile in 5 prepubertal boys (11-13 years old). Via a muscle biopsy, the authors observed an increase in the concentration of the oxidative enzyme, succinicate dehydrogenase (SDH) of 30% post training. Work by the same authors in adults, reported that endurance athletes show greater levels of SDH (50% higher) when compared to sedentary people (Eriksson et al., 1973). Moreover, it has been reported that the concentration of another aerobic enzyme, isocitric dehydrogenase (ICDH), is higher in children than typically found in adults and also that there is a lower ratio of ICDH to PFK (Haralambie, 1982), suggesting that children might be preferentially adapted to aerobic metabolism. Arguably this may leave less scope to improve the aerobic enzymatic profile of the muscle with training potentially explaining the blunted training effect in children, but with such little data derived from limited numbers of children this has to remain speculation.

Aerobic fitness: A blunted response or too narrow view of the concept?
From the presented data it is clear that children can experience improvements in cardiorespiratory function after exercise training, although not to the same extent as adolescents or adults. The majority of the studies report increments in peak VO₂ typically around 6%, with exceptional cases when studies demonstrate improvements greater than 15%. However methodological flaws related with the intensity used during the training study, initial level of fitness of the subjects, lack of controls, a mismatch of training type with testing mode (e.g. swim training with cycle testing), insufficient training intensities, all possibly contribute to the reduced aerobic fitness gains.

Another issue related with studying aerobic fitness may be related with to the over-reliance on using peak VO₂ as a 'gold standard'. What if the concept of aerobic fitness has not been interpreted fully in children? Does focusing on changing peak VO₂ provide too narrow a focus for what is meant by the term 'aerobic trainability'? After all peak VO₂ is only one component of the factors that contribute to endurance exercise performance.

According to Bosh, 2006, misunderstandings surrounding scientific dogma may have influenced sport scientists' interpretation of the evidence that is used to extrapolate and generalize results. As such, because studies mainly used peak VO₂ as a parameter of evaluation it is hard to ascertain if other changes in aerobic fitness were also present.

Indeed adult data indicates that peak VO₂ is of limited value to athletes, because exercise scientists and coaches cannot use it as a predictor for future performance, or for the prescription of potentially "optimal" training (Bosh, 2006; Jones, 2006). Although peak VO₂ might show limited scope for change, other physiologic parameters known to be related to aerobic exercise performance such as lactate and ventilatory thresholds, exercise economy, and VO₂ kinetic response might have been altered favourably (Jones, 2006; Jones and Carter, 2000). The longitudinal studies on Paula Radcliff (Jones, 2006) and Lance Armstrong (Coyle, 2005) epitomize this; both these world class athletes showed a relatively stable peak VO₂ during long periods of their careers, yet economy and performance enhanced significantly in the same time.

In children the data indicate similar adaptations are evident. Welsman et al. (1997) found no changes in peak VO₂ in prepubertal girls after endurance training; however a significant reduction in submaximal exercise lactate levels were noted in these same children. Haffor et al., 1990 noted an improvement in lactate threshold in a group of 11-year-old boys after interval training yet peak VO₂ was unchanged (Haffor et al., 1990). Studies on young children have also seen improvements in running economy and performance set against no changes in peak VO₂ (Daniels et al., 1978; Krahenbuhl et al., 1989).

Key point summary
Peak VO₂ can significantly improve with training in children by approximately 5%.

Why children show a 'reduced' trainability of peak VO₂ in comparison to adults is still not clear.

Improvements in exercise economy, lactate threshold and performance may occur without any change in peak VO₂.

Sport development practices and structures direct children to specialize in one or two sports from an early age, especially if they seek to aspire to perform at the very highest levels. The "catch them young philosophy" that matched the beliefs of many coaches who think that in order to achieve success at senior level it is necessary to start intensive training well before puberty (Baxter-Jones, 2007; Baxter-Jones and Mundt, 2007), has meant many of our youngsters are training intensively and for considerable hours by the time they become adolescents. This is of great importance as they are often asked to train at levels that could be considered exhaustive even for adults, but also they are assumed to be capable of withstanding the physical and psychological pressures that participation in elite sport can entail (Borms, 1986; Kentta et al., 2001). It is this success driven environment, that potentially causes the athlete to engage in very hard training without being aware if it really is of benefit for his / her own performance (Kentta et al., 2001). In addition to the long hours of intensive, repetitive training and strict dietary regimens, it is also important to consider the impact on the young athlete's socio-cultural life; separations from the family to train or compete away from home, the impact on academic performance and the reduced opportunities to make school friends etc (Coakley, 1992; Hollander, 1995; Kentta et al., 2001; Kentta, 2001; Morgan, 1987). Consequently, the combination of heavy training, inadequate recovery and limited social support networks for the young athlete can result in overtraining in even young and aspiring elite athletes (Coakley, 1992; Kentta et al., 2001). It is therefore important that sport scientists, coaches, medics and parents start to become aware of the potential negative health implications (physical, physiological and psychological) of such training practices in young athletes (American Academy of Pediatrics, 2000).

Overtraining concept
According to Kreider et al. 's (1998) definition of overtraining, this state results from an accumulation of training and non-training stressors that has a detrimental long-term effect on performance, with a recovery period that may take several weeks or months. Overreaching, which is considered to be a normal process of training, is defined as an accumulation of training and non-training stressors that lead to a short-term decrease in performance, which can be overcome with recovery lasting days or a few weeks (Kreider et al., 1998). However overreaching may develop into overtraining if sufficient recovery time is not given, indicating that overreaching and overtraining are just two ends of the same continuum.

Overtraining is fundamentally an imbalance between training fatigue and non-training stressors, and recovery. Furthermore, it is associated with a variety of symptoms that often vary considerably across individuals (Kentta et al., 2001). The latter has led to disagreements as to which signs and symptoms need to be present for an athlete to be defined as being overtrained or overreached (Meesuen et al., 2006; Uusitalo, 2006). A recent position statement by the European College of Sports Sciences tried to provide clarity to the definition by recommending that overreaching was divided into two different stages, functional overreaching and non-functional overreaching, with the latter eventually leading to the overtraining syndrome (Meesuen et al., 2006). However, the statement did not provide guidance over the duration of each stage (Uusitalo, 2006), making it hard to distinguish which state is the athlete in and also, when does functional overreaching turn into non-functional overreaching and ultimately overtraining? With disagreement over not only the presenting symptoms of but also the definition of overreaching and overtraining, the simplicity of the categorizations provided by Kreider et al., 1998 seems more useful to employ.

Overtraining in adults and young athletes
Overtraining syndrome has been extensively studied in adult athletes, and therefore data on its prevalence are known. Morgan and colleagues (1988a) reported that approximately 10% of male collegiate swimmers indicated being overtrained, although rates of up to 21% have been noted in other swimmers (Hooper et al., 1995; O'Connor et al., 1989). The signs and symptoms of overtraining appeared in more than 50% of professional soccer players during a 5-month competitive season (Lehmann et al., 1992), and 33% in basketball players during a training camp of 4 weeks (Verma et al., 1992). In two separate research studies in elite US distance runners, Morgan et al. found that 60% of women and 64% of men reported being overtrained at some point in their career (Morgan et al., 1988b; Morgan et al., 1987b). Importantly, the rate of overtraining occurrence dropped to 33% when non-elite women runners were considered (Morgan et al., 1987b), which is indicative of the close relationship between performance level and the prevalence of overtraining.

While overtraining is acknowledged to be of concern for athletes both in endurance and non-endurance sports (Kentta et al., 2001; Kreider et al., 1998), less data are known regarding incidence rates in individual sports compared to team sports. The few studies available report incidence ranging from 10-64% in individual sports and between 33-50% in team sports (Hooper et al., 1995; Morgan et al., 1988a; O'Connor et al., 1989).

Is overreaching and overtraining prevalent in young athletes?
Very few investigations have been directed to young athletes and therefore, it is not known whether similar responses to chronic over-exercising in adults are evident in children. Recently a study involving adolescent swimmers (13-18 year old) looked at the prevalence of overtraining across different countries (Japan, USA, Sweden and Greece) and found that 35% had been overtrained at least once (Raglin et al., 2000). Kentta and colleagues (2001) observed higher incidence rates for individual sports (48%) compared with team sports (30%) and less physically demanding sports (18%). In addition, in 1992, Coakley identified 15 adolescents who had been overtrained during their careers (ages ranged from 15 to 19 years); the author concluded overtraining was grounded in the social organization of sport and recommended changes in the organization of high performance sports; Also psychodoping was acknowledged as a social problem that restricts athletes from creating a "healthy" identity due to their lack of control over their own lives (Coakley, 1992).

Signs and symptoms in adult athletes
A considerable number of studies have investigated the common signs and symptoms reported in athletes from a range of different sports. Fry et al.'s (1991) review listed more than 90 different symptoms that are reported by overtrained athletes. Unfortunately, such a long list makes it difficult for a coach or an athlete to know which signs or symptoms to be concerned about, but further complicating the picture is the large interindividual variability found within athletes practicing the same sport (Verma et al., 1992).

One of the main complaints from athletes is a decrease in performance (Budget, 1998), coupled with chronic fatigue and apathy (Armstrong and VanHeest, 2002), which will often be ignored. Nevertheless, the responses of athletes to overtraining have essentially been measured from 2 categories of symptoms: physiological and psychological. Several physiological markers have been proposed as important and potential parameters for the diagnose of overtraining such as: decreased or suppressed levels of immunoglobulin A (IgA) (Gleeson et al., 1999; Gleeson and Pyne, 2000; Ring et al., 2005), decreased salivary testosterone to cortisol ratio (Passelergue and Lac, 1999), increased salivary cortisol (O'Connor et al., 1989), decreased heart rate variability (Hedelin et al., 2000; Pichot et al., 2002), increased catecholamines (Hooper et al., 1995).

However, some of the parameters, like hormonal markers, should not be used as a single measure, since they show a very high variability (Urhausen and Kindermann, 2002). As example, it has been proposed that a decrease in cortisol levels occurs in the more chronic state of overtraining (Lehmann et al., 1998), whereas an increase would represent an acute higher physiological strain (Kirwan et al., 1988).

There is unfortunately no single practical, valid and reliable physiological marker that can be used to enable a clear and quick diagnoses of athletes who are entering this state (Kentta et al., 2001). Nevertheless, some symptoms appear to be more frequently reported than others; such as a higher incidence in infectious disorders (upper respiratory tract infections), loss of appetite, unexpected weight loss, sleep disturbances and emotional disturbances (Armstrong and VanHeest, 2002; Fry et al., 1991; Morgan et al., 1987a; Raglin et al., 2000).

Signs and symptoms in young athletes
Although few investigations have collected data on athletes in regards to overtraining, there is some evidence suggesting that the signs and symptoms in young athletes are typically the same or similar to the ones found in the adult population. The mentioned cross-cultural study performed on swimmers found that the highest complaint from swimmers was primarily an increased perception of effort, followed by feelings of heaviness (Raglin et al., 2000), which resembles the results from another study on young elite athletes (Kentta et al., 2001). Also, swimmers who reported to be unmotivated had higher levels of mood disturbances at all assessment points compared with overtrained swimmers who retained their incentive to train (Raglin et al., 2000). The study by Raglin and colleagues (2000) further reported more symptoms associated with overtraining like muscle soreness, sleep disturbances and loss of appetite. Furthermore, athletes were also asked to report additional symptoms during the overtrained periods; they found these to be essentially of psychosocial nature, i.e. social problems (family, boyfriend / girlfriend, coach or friends), negative feelings like decreased interest in training and competition and frustration to continue training. Finally, many other psychological problems have also been described like athletes' decreased self-confidence and ability to focus, short temper, heightened levels of irritability, depression, sadness, and elevated levels of perceived stress (Kentta et al., 2001).

Since 20 years of research on overtraining indicates that objective physiological markers have generally not proven useful to monitor training and performance, therefore the need to increase athletes' self-awareness, i.e. making them more aware of physical or psychological stress has been highlighted (Kentta and Hassmen, 1998; Kentta et al., 2001).

This latter issue is important as clinical depression has been reported in overtrained athletes (Hooper et al., 1995; Kentta et al., 2006; Morgan et al., 1987a; 1988a. In regard to the psychological factors associated with overtraining in youth sports, the use of the POMS (profile of mood states) scale that assesses relevant emotional shifts in tension, depression, anger, vigour, fatigue and confusion has proven useful. As overtraining progresses the deterioration of positive moods and the increase in negative moods can clearly be tracked by POMS. Furthermore, the necessity to increase the awareness of youth sport coaches about this issue has been stressed elsewhere (Hollander et al., 1995). Likewise, parents can unduly pressurize their child by overemphasizing winning, holding unrealistic expectations, criticizing them and pushing them to play. All of these factors reiterate the need for parents to be educated about their critical role in their child's sporting development (Gould and Eklund, 1991).

Overtraining or depression?
It has been recently recognised that the overtraining signs and symptoms are actually very similar to the ones encountered in the condition traditionally known as depression (Armstrong and VanHeest, 2002). According to the American Psychiatric Association, there are two sub-types of this condition, but concerningly the one that most closely resembles overtraining syndrome is known as major depression (Yudofsky and Hales, 1992). Major depression affects an individual's thoughts, feelings, physical health, behaviour and ability to function on daily activities. Furthermore, it involves at least two weeks of depressed mood or lack of interest on nearly all life activities, with individuals reporting similar signs and symptoms to the ones seen in overtraining - problems with sleep, lack of appetite, irritable mood, loss of body weight, loss of motivation (Yudofsky and Hales, 1992). Finally, overtraining syndrome shares similar brain structures, endocrine pathways and immune responses to the ones reported for major depression (Morgan et al., 1987a; O'Connor et al., 1991). In fact, Morgan and colleagues (1987a) reported that approximately 80% of athletes with overtraining had a psychopathology similar to people with psychological depression.

Moreover, even highly motivated athletes will have difficulties coping with the way they perceive poor performance, because of their desire to perform well and win (Coakley, 1992; Hanna, 1979). One of the reasons for the frustrations seen in these athletes who are performing badly, has to do with the pressure they put on themselves to achieve their goals (Hanna, 1979). The frustration due to the lack of performance typically drives the athlete to further increase the amount of training that he / she is doing, which just exacerbates the situation by increasing physical and emotional fatigue, with a consequent worsening in performance (Kentta et al., 2001). This cycle (process) manifests itself by changes in mood, sleep disturbances, losses in appetite and body weight, and an increase in depressive symptoms, eventually leading to depression (Armstrong and VanHeest, 2002).

Prevention of overtraining
The limited data available suggests overtraining is prevalent in young athletes, therefore finding ways to identify those children with the condition or at risk of becoming overtrained, supported by constructive ways to treat, or better still avoid, overtraining would be invaluable.

Accordingly, it seems appropriate to address this topic by taking a multidisciplinary view (Birrer and Levine, 1987; Borms, 1986; Kentta et al., 2001); we believe that psychobiologic parameterization of training can offer an effective monitoring strategy for child athletes in the detection and prevention of overtraining. Future overtraining research should try to incorporate the collection of emotional parameters questionnaires such as POMS, Training-Induced Distress scale (Raglin and Morgan, 1994), or overtraining questionnaires (Maso et al., 2004) and, at the same time collect objective physiological markers such as immunoglobulins (IgA), hormonal parameters (testosterone and cortisol), heart rate variability etc. By so doing, this will allow us to understand how variations in measures of mood correlate with physiological markers (Kentta and Hassmen, 1998), but also view the problem holistically. Treatment strategies can therefore be focused not just on physiological recuperation but also address the personal identity and emotional issues in parallel.

Key-point summary
The limited available evidence seems to point to an occurrence of overtraining in young athletes around 30% and therefore should be investigated more extensively in the future.

Overtraining and clinical major depression share many similarities, therefore overtraining should be studied by acknowledging this perspective.

Collection of physiological, psychological and emotional parameters may be useful to improve our understanding of overtraining and to provide new perspectives on how to look at this important issue in child athletes.

Children's strength, anaerobic and aerobic power is trainable, although the improvements may be smaller than seen in adults. The underlying mechanisms responsible remain to be conclusively determined, but improvements in non-invasive technologies will aid our ability to investigate these issues for children in the future. Overtraining seems to be present in some young athletes, yet we need more research to allow us to recognize the key signs / symptoms along with identifying the central determining factors to allow us to help prevent this condition arising in the first place.

• Children's strength, anaerobic and aerobic power is trainable, although the improvements may be smaller than seen in adults.
• Children can demonstrate significant gains in muscle strength with resistance training (13 - 30%).
• Improvements in mean power (3 - 10%) and peak power (4 - 20%) are reported in children.
• Aerobic fitness can improve with training in children by approximately 5%.
• Limited available evidence indicates an occurrence of overtraining in young athletes of around 30%.