Abstract

Compulsive exercise is a form of addictive behavior: a rigid, excessive pattern of physical training pursued despite injury, illness, or the loss of the activity's original purpose, and accompanied by distress when it is prevented. Veale's distinction between primary dependence, in which exercise is the end in itself, and secondary dependence, in which it serves the weight and shape control of an eating disorder, still organises the field. The behavioral-addiction account models the disorder with Griffiths' six components — salience, mood modification, tolerance, withdrawal, conflict, and relapse — operationalised by the Exercise Addiction Inventory and the Exercise Dependence Scale. Biological hypotheses implicate endorphin, dopamine, and endocannabinoid signalling, while the dualistic model of passion locates the fault in obsessive rather than harmonious engagement. Three interactive demonstrations trace the components screen, tolerance-driven dose escalation, and the separation of commitment from dependence.

Keywords: exercise dependence, exercise addiction, compulsion, tolerance, behavioral addiction

A runner trains through a stress fracture, misses a family event rather than skip a session, and feels anxious and guilty on a forced rest day. The training that began as a route to fitness has become the organising centre of a life, defended against injury, relationship strain, and the athlete's own stated wish to cut back. What makes this pattern compulsive is not the volume of exercise alone — elite athletes train enormous loads without dependence — but the grammar of the behavior: it is rigid, it is continued despite harm, it is pursued to relieve a worsening unease rather than for its own reward, and skipping it provokes withdrawal (Veale, 1987; Berczik et al., 2012).

For most of the twentieth century excessive exercise was noticed mainly as a symptom of anorexia nervosa. Veale's proposal that exercise could be a dependence in its own right — with the same salience, tolerance, withdrawal, and relapse seen in substance use — reframed it as a candidate behavioral addiction, the drug-free pursuit of an activity that recruits the motivational machinery of addiction (Veale, 1987; Freimuth et al., 2011).

Key Takeaways
  • Compulsive exercise is excessive, rigid training continued despite harm and accompanied by withdrawal when prevented — MeSH files it as a kind of addictive behavior.
  • Primary dependence treats exercise as the end itself; secondary dependence is driven by the weight and shape control of an eating disorder. The two have different prognoses and are distinguished at assessment.
  • The behavioral-addiction account scores six components — salience, mood modification, tolerance, withdrawal, conflict, relapse — with the Exercise Addiction Inventory and Exercise Dependence Scale.
  • Tolerance drives escalation: the training dose needed for the same mood effect grows over time, pushing volume toward the unsustainable and the injurious.
  • The disorder is defined by obsessive rather than harmonious engagement; high training volume with healthy commitment is not exercise addiction.

What Compulsive Exercise Is

Compulsive exercise — also called exercise dependence, exercise addiction, or obligatory exercise — is defined by the same break with the ordinary logic of reward that defines compulsion generally. Ordinary training is performed for a valued outcome — fitness, competition, enjoyment — and is adjusted or abandoned when the outcome changes, as when injury makes a session counterproductive. Compulsive exercise breaks this contingency: it is continued through injury and illness, maintained when it no longer serves any athletic goal, and experienced as difficult to control. Its diagnostic weight rests not on the amount of exercise but on this rigidity and on the harm tolerated to sustain it (Veale, 1987; Berczik et al., 2012).

The field's oldest and most durable distinction is Veale's division into primary and secondary exercise dependence. In primary dependence, exercise is the object of the dependence itself: the person trains to obtain the exercise-related mood effect and to avoid the dysphoria of not training, with no other disorder driving the behavior. In secondary dependence, excessive exercise is instrumental to the weight and shape concerns of an eating disorder, and is better understood as one of that disorder's compensatory behaviors. The two are separated at assessment because they carry different risks and call for different treatment, and because secondary cases — entangled with disordered eating — are both more common and more dangerous (Veale, 1987; Cunningham et al., 2016).

The relationship with eating disorders is close enough that the most widely used clinical instrument, the Compulsive Exercise Test, was developed explicitly within the eating-disorder literature to capture the cognitive and affective features of driven exercise — avoidance of negative affect, rule-bound rigidity, and exercise for weight control — rather than volume alone (Taranis et al., 2011; Lichtenstein et al., 2014).

Measuring Exercise Addiction

Because exercise is healthy in most doses, the measurement problem is to separate a disorder from an intense but benign commitment. The dominant solution adapts Griffiths' general components model of addiction — the proposal that every addiction, substance or behavioral, expresses six common components — to exercise. The six are salience (exercise dominates thinking and daily life), mood modification (it is used to alter affect), tolerance (increasing amounts are needed for the same effect), withdrawal (unpleasant feelings when it is prevented), conflict (it damages relationships, work, or health), and relapse (a return to excess after cutting back) (Griffiths, 2005).

Figure 1

The Six-Component Model of Exercise Addiction

A hub-and-spoke diagram of the six components of exercise addiction A central node labelled Exercise addiction connects by six spokes to six surrounding nodes: salience, mood modification, tolerance, withdrawal, conflict, and relapse. Exercise addiction Salience Mood modification Tolerance Withdrawal Conflict Relapse
Note. Each component is scored by the screening instruments described below; a person is flagged when enough components are endorsed at sufficient intensity. Adapted from the components model of addiction (Griffiths, 2005). Demo 1 scores these six components interactively.

Two brief screens operationalise this directly. The Exercise Addiction Inventory poses one item per component on a five-point scale, summing to a score from 6 to 30, with a cutoff of 24 flagging someone at risk of exercise addiction (Terry et al., 2004). The Exercise Dependence Scale, built on the DSM criteria for substance dependence, yields both a continuous score and a three-way classification into dependent, symptomatic, and asymptomatic, and distinguishes physiological from nonphysiological dependence (Hausenblas & Symons Downs, 2002). A head-to-head psychometric comparison found both instruments valid and well correlated, with the shorter inventory trading a little information for speed (Mónok et al., 2012). Demo 1 scores the six-item inventory against its cutoff.

Table 1

Three Validated Screens for Compulsive Exercise

Instrument Form and scale Output Primary focus
Exercise Addiction Inventory (EAI) Six items, one per component, each on a five-point agreement scale Sum from 6 to 30; a score of 24 or above flags someone at risk A brief general screen for the six-component pattern
Exercise Dependence Scale (EDS) Multi-item, mapped to the DSM criteria for substance dependence A continuous score plus a dependent, symptomatic, or asymptomatic classification Multidimensional dependence, separating physiological from nonphysiological
Compulsive Exercise Test (CET) Multi-item, with cognitive and affective subscales A continuous score across subscales such as avoidance and rule-driven behavior Driven exercise within eating disorders, not volume alone

Screening with the Exercise Addiction Inventory

The six-item Exercise Addiction Inventory scores one statement for each of Griffiths’ components of behavioural addiction on a 1–5 scale. The sum runs 6–30, and a total of 24 or more flags someone as at risk of exercise addiction. Set each item as it fits and watch the total cross the cut-off — a screen, never a diagnosis.

Salience: Exercise is the most important thing in my life.3/5
Conflict: Conflicts have arisen with my family or partner about my exercise.3/5
Mood modification: I use exercise as a way of changing my mood.3/5
Tolerance: Over time I have increased the amount of exercise I do in a day.3/5
Withdrawal: If I skip a session I become moody or irritable.3/5
Relapse: If I cut down and then start again, I return to the same amount.3/5
cut-off 24630
Total score 18/30. This is below the cut-off of 24: committed exercise can score highly on salience and mood modification without reaching the at-risk band. It is the pattern of conflict, withdrawal and escalation together that distinguishes dependence.

Why Exercise Becomes Compulsive

Two complementary kinds of explanation sit beneath the screening. One is psychological and continuous with the general theory of addiction; the other is biological, concerned with what exercise does to the brain's reward chemistry. Neither alone accounts for the disorder, and both must explain the same central fact: that the great majority of committed exercisers never become dependent (Weinstein & Weinstein, 2014; Colledge et al., 2020).

Tolerance, withdrawal, and escalation

The engine the components model shares with substance addiction is the tolerance–withdrawal pair. Exercise reliably modifies mood — the acute improvement in affect after a bout is among the most robust findings in exercise psychology — and a person can come to rely on that effect to regulate a baseline unease. With repetition the effect of a fixed dose declines, so that more training is needed to reach the same relief; meanwhile the unease on rest days deepens, because the regulated baseline has drifted downward. The result is escalation: training volume grows not toward a performance goal but to chase a receding mood benefit and to stave off a worsening withdrawal, a trajectory that pushes the exerciser toward overuse injury and exhaustion (Freimuth et al., 2011; Berczik et al., 2012). The worked example quantifies this dose growth, and Demo 2 runs the escalation interactively.

Tolerance forces the dose to escalate

Suppose a fixed daily dose of exercise once delivered enough relief, but each week tolerance erodes its effect by a fraction t. To hold the same relief, the dose must grow geometrically: dn = 60 · (1 / (1 − t))n minutes. Raise the tolerance rate and watch how fast the training load climbs toward a life ceiling of 120 min/day — the point at which exercise crowds out work, rest and relationships.

Weekly tolerance erosion (t)0.10
060120180240life ceiling 120 mindose (min/day)week (n)week 6.58
With t = 0.10 the weekly multiplier is 1.111: the dose runs 102 min by week 5 and 172 min by week 10. It reaches the 120-minute life ceiling at week 6.58. (This is the article's worked example.)

The biology of the exercise reward

What makes exercise capable of this is a set of neurochemical effects it shares, in muted form, with drugs of abuse. The oldest hypothesis is the endorphin account — that exercise releases endogenous opioids producing the analgesia and euphoria of the runner's high. Evidence has since widened to the dopamine reward system, implicated in the incentive salience exercise and its cues acquire, and to the endocannabinoid system, whose exercise-induced activation now has better support than the opioid story for the acute mood effect. None of these is sufficient on its own, and the honest summary is that exercise engages overlapping reward pathways strongly enough to support dependence in vulnerable individuals, without any single transmitter being the cause (Weinstein & Weinstein, 2014).

Commitment versus dependence

The hardest discrimination is between a dependent exerciser and a merely devoted one, because both train hard and value training highly. The dualistic model of passion draws the line through why the person is engaged (Vallerand et al., 2003). Harmonious passion is an autonomous love of the activity that remains under the person's control and coexists with the rest of life; obsessive passion is a controlled, contingent attachment in which the person feels compelled to exercise and other domains suffer. In athletes and leisure exercisers alike, it is obsessive passion — not training volume, and not harmonious passion — that predicts exercise addiction, which is why a high-volume athlete with harmonious passion is not a case (de la Vega et al., 2016; Di Lodovico et al., 2019). Demo 3 separates the two passions and the risk each carries.

Committed or dependent? Two kinds of passion

The dualistic model separates two ways a valued activity is internalized. Harmonious passion is taken up freely and sits alongside the rest of life; obsessive passion controls the person, who feels compelled to exercise even when it conflicts with everything else. The same high training volume can come from either. Set each and read where the behaviour falls on the committed–dependent spectrum.

Harmonious passion (chosen, integrated)0.60
Obsessive passion (compelled, controlling)0.40
obsessive-dominantharmonious-dominantharmonious passion →obsessive passion →
Risk index 0.39 — this profile reads as mixed. The two passions are balanced: high commitment shading toward control. Watch for the shift — it is a rising ratio of obsessive to harmonious passion, not volume alone, that marks the slide into dependence.

Worked Example

Consider the tolerance account and the escalation it predicts. Suppose a person exercises to obtain a fixed mood benefit of T = 30 units. The mood return per minute of training — the sensitivity — starts at s0 = 0.5 units/min, so the dose that delivers the target benefit is d0 = T / s0 = 60 minutes. Tolerance erodes sensitivity by a fixed fraction r = 0.10 each week, so after n weeks sn = s0(1 − r)n, and the dose required to still reach T is

dn = T / sn = d0 (1 − r)−n = 60 × (1 / 0.9)n

Then:

- Week 0: d0 = 60 min - Week 5: d5 = 60 × (1.1111)5 = 60 × 1.6935 ≈ 101.6 min - Week 10: d10 = 60 × (1.1111)10 = 60 × 2.8680 ≈ 172.1 min

The required dose grows geometrically. Suppose the largest daily training the person can sustain without breakdown is a ceiling of d = 120 minutes. The escalation reaches that ceiling when 60 × (1.1111)n = 120, i.e. (1.1111)n = 2:

n = ln(2) / ln(1.1111) = 0.6931 / 0.10536 ≈ 6.58 weeks

So in under seven weeks the dose needed merely to maintain the original mood benefit exceeds what the body can absorb — the point at which continued escalation produces overuse injury rather than benefit. This is the compulsive trap in miniature: the behavior must grow to keep delivering a constant, and shrinking, return, and the growth itself becomes the harm. Demo 2 traces dn across weeks and marks this crossover.

Discussion

The psychological and biological accounts are read as complementary. The components model and the tolerance–withdrawal engine explain the form of the disorder — why it escalates, why rest is aversive, why it persists against harm — while the neurochemistry of the exercise reward explains how an activity with no ingested substance can nonetheless sustain a dependence. The passion model supplies the discrimination the others lack, separating the dependent exerciser from the devoted one by the quality of engagement rather than its quantity (de la Vega et al., 2016; Weinstein & Weinstein, 2014).

The central unresolved question is nosological. Exercise addiction appears in no major diagnostic manual: it is not in the DSM-5, and ICD-11 recognises it nowhere as such. The prevalence of genuine dependence in the general population is low — on the order of a few percent, and far lower than self-identified addiction surveys suggest — which keeps alive the worry that the construct over-pathologises healthy dedication (Mónok et al., 2012; Colledge et al., 2020). The strongest case for the disorder's reality is the secondary form's tight coupling to eating disorders, where driven exercise is unambiguously harmful; the primary form, exercise dependence with no other disorder, remains the more contested category (Cunningham et al., 2016).

Current Directions

Several fronts are active. The first is a sustained effort to pin down prevalence and risk: meta-analytic work finds the highest risk concentrated in endurance and aesthetic sports and in those with disordered eating, sharpening the picture of who is vulnerable rather than treating exercisers as uniform (Di Lodovico et al., 2019; Colledge et al., 2020). The second is measurement refinement — reconciling the several instruments, separating primary from secondary cases at screening, and distinguishing the obsessive engagement that signals a disorder from the harmonious engagement that does not (Mónok et al., 2012; de la Vega et al., 2016). The third is the search for mechanism, as the endocannabinoid account displaces the older endorphin story and neuroimaging begins to test whether exercise recruits the same incentive-salience circuitry as substance addiction (Weinstein & Weinstein, 2014). Whether exercise addiction will eventually be recognised as a formal diagnosis, and on what evidence, is the question that frames the rest.

Common Misconceptions

Anyone who exercises a lot is addicted.
Volume alone does not define the disorder. Elite athletes sustain very high training loads with harmonious passion and no dependence. The marks of compulsive exercise are continuation despite harm, withdrawal when prevented, and obsessive rather than harmonious engagement (de la Vega et al., 2016).
Exercise addiction is just a figure of speech.
It meets the functional criteria of a behavioral addiction — salience, tolerance, withdrawal, conflict, relapse — and in its secondary form is tightly bound to eating disorders. It is a candidate disorder with validated screening instruments, even though no major manual yet lists it (Griffiths, 2005; Terry et al., 2004).
Primary and secondary dependence are the same thing.
They are separated deliberately. In primary dependence exercise is the end itself; in secondary dependence it serves the weight control of an eating disorder and is a compensatory behavior of that disorder. The distinction changes both prognosis and treatment (Veale, 1987; Cunningham et al., 2016).
The runner's high is pure endorphins.
The endorphin account is the oldest but no longer the best supported. Exercise-induced activation of the endocannabinoid system has stronger evidence for the acute mood effect, and the dopamine reward system is implicated in the incentive pull of training. No single transmitter is the cause (Weinstein & Weinstein, 2014).

Glossary

Behavioral addiction.
An addiction to an activity rather than an ingested substance, proposed to share the motivational components of drug addiction; compulsive exercise is a candidate case.
Components model.
Griffiths' proposal that every addiction expresses six common components — salience, mood modification, tolerance, withdrawal, conflict, and relapse — the framework the exercise screens operationalise.
Compulsive Exercise Test (CET).
A questionnaire developed within the eating-disorder literature to capture the cognitive and affective features of driven exercise — avoidance of negative affect, rule-bound rigidity, weight-control motives — rather than training volume.
Dualistic model of passion.
Vallerand's framework distinguishing harmonious from obsessive passion for an activity; applied to exercise, it locates the disorder in obsessive rather than harmonious engagement.
Endocannabinoid system.
A lipid-signalling system activated by sustained exercise, now better supported than the endorphin account for the acute mood lift of the runner's high.
Exercise Addiction Inventory (EAI).
A six-item screen, one item per component of the model, scored 6–30, with a cutoff of 24 identifying someone at risk of exercise addiction.
Exercise Dependence Scale (EDS).
A multidimensional instrument built on the DSM substance-dependence criteria, yielding a continuous score and a classification into dependent, symptomatic, and asymptomatic.
Harmonious passion.
An autonomous, controllable love of an activity that coexists with the rest of life; in the dualistic model it does not predict exercise addiction even at high training volume.
Obsessive passion.
A controlled, contingent attachment in which a person feels compelled to engage and other life domains suffer; the form of passion that predicts exercise addiction.
Primary exercise dependence.
Dependence in which exercise is the end in itself, pursued for its mood effect and to avoid withdrawal, with no other disorder driving it.
Runner's high.
The acute euphoria and analgesia reported after sustained exercise; attributed historically to endorphins and more recently to endocannabinoid signalling.
Secondary exercise dependence.
Excessive exercise that serves the weight and shape control of an eating disorder, understood as a compensatory behavior of that disorder rather than a dependence in its own right.
Tolerance.
The declining effect of a fixed dose with repetition, so that ever more exercise is needed for the same mood benefit; the driver of escalation in the worked example.
Withdrawal.
The aversive affective state — anxiety, irritability, guilt, low mood — that follows prevented or missed exercise in a dependent person, and whose relief motivates further training.

Key Researchers

Zsolt Demetrovics

(Institute of Psychology, Eötvös Loránd University). A behavioral-addictions researcher who has led large psychometric comparisons of the exercise-addiction instruments and epidemiological reviews of the disorder. ORCID · Wikipedia

Danielle Symons Downs

(Department of Kinesiology, The Pennsylvania State University). Co-developed the Exercise Dependence Scale and its multidimensional model of physiological and nonphysiological dependence. Faculty · Google Scholar

Mark D. Griffiths

(International Gaming Research Unit, Nottingham Trent University). Formulated the six-component model of addiction that underlies the exercise screens and co-developed the Exercise Addiction Inventory. ORCID · Wikipedia

Heather A. Hausenblas

(School of Applied Health Sciences, Jacksonville University). Co-developed the Exercise Dependence Scale and its systematic grounding in the substance-dependence criteria. Faculty

Caroline Meyer

(University of Warwick). Co-developed the Compulsive Exercise Test, bringing the cognitive and affective features of driven exercise into eating-disorder assessment. Faculty · Google Scholar

Attila Szabo

(Institute of Health Promotion and Sport Sciences, Eötvös Loránd University). A leading researcher on exercise addiction, co-author of the Exercise Addiction Inventory and of work separating commitment from dependence. Faculty · Google Scholar

David Veale

(Institute of Psychiatry, Psychology & Neuroscience, King's College London). Introduced the distinction between primary and secondary exercise dependence that still organises the field. ORCID · Wikipedia

Frequently Asked Questions

When does exercise become compulsive rather than just dedicated?

The line is not drawn at a number of hours. Exercise is compulsive when it is continued despite injury, illness, or clear harm to relationships and work; when missing it provokes anxiety, guilt, or low mood; and when the person feels driven rather than freely choosing. Dedicated high-volume training without these features is not a disorder.

What is the difference between primary and secondary exercise dependence?

In primary dependence, exercise is the end in itself, pursued for its mood effect and to avoid withdrawal, with no other disorder behind it. In secondary dependence, excessive exercise serves the weight and shape control of an eating disorder and is one of its compensatory behaviors. Secondary cases are more common and are assessed and treated within eating-disorder care.

How is exercise addiction measured?

With brief validated questionnaires. The Exercise Addiction Inventory scores six components on a 6-to-30 scale with a cutoff of 24 for being at risk. The Exercise Dependence Scale, built on the substance-dependence criteria, classifies people as dependent, symptomatic, or asymptomatic. The Compulsive Exercise Test captures the cognitive side, especially in eating disorders.

Why does the amount of exercise keep increasing?

Through tolerance. Exercise reliably lifts mood, but the effect of a fixed dose fades with repetition, so more training is needed to reach the same benefit, while rest days become more aversive as the baseline mood drifts down. Volume then grows to chase a receding effect and escape a worsening withdrawal, pushing toward overuse injury.

Is the runner's high caused by endorphins?

Partly, but the endorphin account is the oldest rather than the best supported. Exercise activates the endocannabinoid system, which now has stronger evidence for the acute mood effect, and engages the dopamine reward system that gives training its incentive pull. The high reflects several overlapping systems, not one.

Can someone be a committed athlete without being addicted?

Yes, and this is the key discrimination. The dualistic model of passion distinguishes harmonious passion (an autonomous love of the activity under the person's control) from obsessive passion, a compelled, contingent attachment. Only obsessive passion predicts exercise addiction; harmonious passion does not, even at very high training volume.

Is exercise addiction an official diagnosis?

No major diagnostic manual lists it. It is absent from the DSM-5 and is not recognised as such in ICD-11. It is a well-studied candidate behavioral addiction with validated screening tools, but its status, especially the primary form with no accompanying eating disorder, remains debated.

Is compulsive exercise a kind of addictive behavior?

Yes. MeSH files compulsive exercise as a narrower category of addictive behavior, and the components model treats it as a behavioral addiction that shares salience, tolerance, withdrawal, conflict, and relapse with substance use: the same functional signature, with an activity in place of a drug.

References

Berczik, K., Szabó, A., Griffiths, M. D., Kurimay, T., Kun, B., Urbán, R., & Demetrovics, Z. (2012). Exercise addiction: Symptoms, diagnosis, epidemiology, and etiology. Substance Use & Misuse, 47(4), 403–417. https://doi.org/10.3109/10826084.2011.639120

Colledge, F., Cody, R., Buchner, U. G., Schmidt, A., Pühse, U., Gerber, M., Wiesbeck, G., Lang, U. E., & Walter, M. (2020). Excessive exercise—A meta-review. Frontiers in Psychiatry, 11, 521572. https://doi.org/10.3389/fpsyt.2020.521572

Cunningham, H. E., Pearman, S., III, & Brewerton, T. D. (2016). Conceptualizing primary and secondary pathological exercise using available measures of excessive exercise. International Journal of Eating Disorders, 49(8), 778–792. https://doi.org/10.1002/eat.22551

de la Vega, R., Parastatidou, I. S., Ruíz-Barquín, R., & Szabo, A. (2016). Exercise addiction in athletes and leisure exercisers: The moderating role of passion. Journal of Behavioral Addictions, 5(2), 325–331. https://doi.org/10.1556/2006.5.2016.043

Di Lodovico, L., Poulnais, S., & Gorwood, P. (2019). Which sports are more at risk of physical exercise addiction: A systematic review. Addictive Behaviors, 93, 257–262. https://doi.org/10.1016/j.addbeh.2018.12.030

Freimuth, M., Moniz, S., & Kim, S. R. (2011). Clarifying exercise addiction: Differential diagnosis, co-occurring disorders, and phases of addiction. International Journal of Environmental Research and Public Health, 8(10), 4069–4081. https://doi.org/10.3390/ijerph8104069

Griffiths, M. (2005). A ‘components’ model of addiction within a biopsychosocial framework. Journal of Substance Use, 10(4), 191–197. https://doi.org/10.1080/14659890500114359

Hausenblas, H. A., & Symons Downs, D. (2002). How much is too much? The development and validation of the Exercise Dependence Scale. Psychology & Health, 17(4), 387–404. https://doi.org/10.1080/0887044022000004894

Lichtenstein, M. B., Christiansen, E., Elklit, A., Bilenberg, N., & Støving, R. K. (2014). Exercise addiction: A study of eating disorder symptoms, quality of life, personality traits and attachment styles. Psychiatry Research, 215(2), 410–416. https://doi.org/10.1016/j.psychres.2013.11.010

Mónok, K., Berczik, K., Urbán, R., Szabo, A., Griffiths, M. D., Farkas, J., Magi, A., Eisinger, A., Kurimay, T., Kökönyei, G., Kun, B., Paksi, B., & Demetrovics, Z. (2012). Psychometric properties and concurrent validity of two exercise addiction measures: A population wide study. Psychology of Sport and Exercise, 13(6), 739-746. https://doi.org/10.1016/j.psychsport.2012.06.003

Taranis, L., Touyz, S., & Meyer, C. (2011). Disordered eating and exercise: Development and preliminary validation of the Compulsive Exercise Test (CET). European Eating Disorders Review, 19(3), 256–268. https://doi.org/10.1002/erv.1108

Terry, A., Szabo, A., & Griffiths, M. (2004). The Exercise Addiction Inventory: A new brief screening tool. Addiction Research & Theory, 12(5), 489–499. https://doi.org/10.1080/16066350310001637363

Vallerand, R. J., Blanchard, C., Mageau, G. A., Koestner, R., Ratelle, C., Léonard, M., Gagné, M., & Marsolais, J. (2003). Les passions de l'âme: On obsessive and harmonious passion. Journal of Personality and Social Psychology, 85(4), 756–767. https://doi.org/10.1037/0022-3514.85.4.756

Veale, D. M. W. (1987). Exercise dependence. British Journal of Addiction, 82(7), 735–740. https://doi.org/10.1111/j.1360-0443.1987.tb01539.x

Weinstein, A., & Weinstein, Y. (2014). Exercise addiction: Diagnosis, bio-psychological mechanisms and treatment issues. Current Pharmaceutical Design, 20(25), 4062–4069. https://doi.org/10.2174/13816128113199990614