It is well established that any level of physical activity can help prevent and treat depression, with more strenuous activity having a greater effect. Understanding the mechanisms driving this antidepressant effect is important because it could allow exercise programmes to be made more effective, accessible, and targeted. Such knowledge could contribute to social prescribing, increasingly a priority for mental healthcare. Importantly, physical activity is highly scalable, low cost, well suited to early intervention, and has beneficial impacts on physical health co-morbidities. This trial may provide initial indications of whether there are sub-groups of depressed individuals who are particularly likely to benefit from physical activity, lead to strategies to personalise physical activity prescription based on motivational factors, and pave the way for augmentative approaches, for example combining physical activity with psychological interventions.

To date the mechanisms driving the antidepressant effects of physical activity in humans are poorly understood. Building on links between depressive symptoms, reward processing and dopamine, plus evidence from animal studies that physical activity is anti-inflammatory and boosts both dopamine and reward processing, the overarching aim of this trial is to understand the mechanisms underlying the effects of physical activity in depression, focusing on the concept of motivation.

The key objective is to conduct a randomised controlled trial (RCT) in N=250 depressed participants comparing aerobic exercise to a stretching/relaxation control condition, examining a range of mechanistic factors. The proposed trial will examine the impact of physical activity at multiple, linked potential levels of explanation: (1) immune-metabolic markers; (2) dopamine synthesis capacity; (3) activation in the brain's reward and effort processing circuitry;(4) effort-based decision making incorporating computational analysis; and (5) symptom networks based on fine-grained, daily measurements.

The primary objective is to conduct a randomised controlled trial (RCT) in N=250 depressed participants comparing aerobic exercise to a stretching/relaxation control condition, examining effects on a range of potential clinical and mechanistic factors: depressive symptoms; immune-metabolic function; activation in the brain's reward and effort processing circuitry using functional magnetic resonance imaging (fMRI); cognitive tasks, focusing on reward processing; and a subset (approximately one-third) of participants will complete L-6-[18F] fluoro-3,4-dihydroxyphenylalnine (18F-DOPA) positron emission tomography (PET).

The secondary objectives are to assess: (1) the degree to which changes in the mechanistic factors are related to changes in interest-activity symptoms of depression resulting from aerobic exercise; (2) whether baseline mechanistic or clinical factors are associated with symptomatic improvement measured by symptom questionnaires following the exercise intervention; (3) whether aerobic exercise-induced changes in the brain circuits underlying cognitive control overlap with those implicated in motivation.

The trial will use an RCT design, with depressed participants randomised to eight weeks of either 45 minutes aerobic exercise of moderate-to-vigorous intensity activity (experimental group: three times per week, N=125) or 45 minutes of non-aerobic stretching/guided relaxation (control group: three times per week, N=125). The target sample size following expected attrition is N~105 per arm. Participants will complete the trial in staggered cohorts, with no more than six participants per class.

Blood and saliva samples will be taken before the intervention at baseline (between weeks -1 and 0), mid-intervention (week 3 and week 4), and post-intervention (week 9 to week 14) visits, to assess changes in immune-metabolic markers. Blood and saliva samples will also be collected at baseline and post-intervention from approximately 30 healthy controls.

Functional neuroimaging during effort-based decision-making and cognitive control will be taken at baseline and post-intervention. The same functional neuroimaging measures will also be collected at baseline and post-intervention from approximately 30 healthy controls.

Cognitive assessments will be completed online at baseline, every other week during the intervention (week 1, week 3, week 5, week 7), and post-intervention. The same cognitive assessments will also be collected at identical time-windows from approximately 30 healthy controls.

Questionnaire assessments will be completed online at baseline, every other week (week 2, week 4, week 6) and post-intervention (week 9 to 14). The same questionnaire assessments will be collected at baseline and post-intervention from approximately 30 healthy controls.

Accelerometers will measure physical activity continuously at baseline and during the intervention. Fitness testing will provide a measure of cardiovascular fitness at baseline and post-intervention visits. Daily depressive symptoms will be recorded using abbreviated scales using the Neureka smartphone app throughout the intervention. Three-monthly follow up of symptoms/cognition from baseline over six months will assess the durability of effects (week 21 and week 33).

In a subset of participants, approximately one-third of the participants will also complete a positron emission tomography (PET) scan pre-randomisation and at a visit during weeks 4-9 to 14, to assess dopamine synthesis capacity. The same PET scan will also be collected from approximately 30 healthy controls.

Inclusion Criteria will include:

1. PHQ9≥12 (moderate depression).
2. Current physical activity level below 30 min moderate physical activity, once per week.
3. Fluency in English.
4. Willingness to undergo the interventions.
5. Age 18-60.
6. Willing and able to provide written informed consent.

Exclusion Criteria will include:

1. Medical contraindications to either intervention.
2. Neurological illness.
3. Past or current diagnosis of psychosis, bipolar disorder, or substance/alcohol use disorder, unless restricted to a depressive episode.
4. Unable to complete self-administered cognitive or questionnaire assessments.
5. Symptoms or cognitive impairment that would limit capacity to consent.
6. Pregnancy.
7. Regular use of anti-inflammatory medication (more than once per week).

We will share de-identified questionnaire, cognitive, activity, EMA and immune-metabolic data through online repositories. We will share de-identified MRI and PET data on request through managed access agreements.

Anonymised data will be available for sharing after publication of the trial results. Data will be available indefinitely.

The control arm is stretching and relaxation. Prior research indicates that participants have similar efficacy expectations regarding mild stretching and aerobic exercise, which will help to improve masking. Due to the nature of the interventions, it is not possible for the participants to be masked as to the intervention they are undergoing, but all study information will be agnostic about which intervention we expect to be more effective, and all analysis will be performed by blinded staff. The trial will assess expectations and preferences about the intervention conditions, allowing for sensitivity analyses.