The decision this page answers. HIIE is promoted as a time-efficient route to fitness (Measuring and Raising Cardiorespiratory Fitness, HIIT vs Continuous Training for Type 2 Diabetes). The standing objection is that it feels worse, and an exercise not done has no effect (the schema’s adherence is part of the effect rule). So the question is a substitution one — HIIE versus continuous moderate exercise (MICE), the realistic alternative — judged on in-task affect, not on whether to exercise at all. This page holds the affective evidence; whether the worse feeling actually lowers uptake is a separate, still-open endpoint (below).
What the source is. Niven 2020 is a gold systematic review and meta-analysis of 33 studies of acute affective responses, comparing HIIE against MICE and against vigorous-intensity continuous exercise (VICE), with affect pooled by the single-item 10-point Feeling Scale (FS; Hardy & Rejeski 1989) at matched time points — during (~50% duration), end, immediately post, post+ (>2 min after), and the lowest value reached (Niven et al., 2020).
The effect table — HIIE minus the continuous comparator
A negative affect MD favours the comparator (HIIE felt less pleasant); a positive arousal MD means HIIE was more activated; a positive enjoyment MD favours HIIE. Affect and arousal are FS/FAS points (roughly an 11-point bipolar scale); enjoyment is on a separate 0-100-type scale, so its magnitude is not comparable to the affect points (Niven et al., 2020).
| Outcome | Comparison | Time point | MD (95% CI) | Read |
|---|---|---|---|---|
| Affect (FS) | HIIE vs MICE | during (~50%) | -1.07 [-1.47, -0.67] | favours MICE |
| Affect (FS) | HIIE vs MICE | end (100%) | -0.72 [-1.64, 0.20] | no difference |
| Affect (FS) | HIIE vs MICE | post | -1.07 [-1.44, -0.70] | favours MICE |
| Affect (FS) | HIIE vs MICE | post+ (5-20 min) | -0.22 [-0.40, -0.04] | favours MICE (small) |
| Affect (FS) | HIIE vs MICE | lowest value | -1.18 [-1.96, -0.41] | favours MICE |
| Affect (FS) | HIIE vs VICE | all five points | 0.13 to 0.95, all NS | no difference |
| Arousal (FAS) | HIIE vs MICE | during (~50%) | +1.08 [0.71, 1.44] | higher in HIIE |
| Arousal (FAS) | HIIE vs MICE | end | +1.68 [1.11, 2.25] | higher in HIIE |
| Arousal (FAS) | HIIE vs MICE | post | +0.62 [0.23, 1.00] | higher in HIIE |
| Arousal (FAS) | HIIE vs MICE | post+ | +0.24 [-0.05, 0.54] | no difference |
| Enjoyment | HIIE vs MICE | post | +6.73 [2.45, 11.01] | favours HIIE |
All analyses carried a large degree of heterogeneity (Niven et al., 2020). HIIE-vs-VICE affect was non-significant at every one of the five time points (50%: 0.33 [-0.46, 1.12]; end: 0.36 [-0.99, 1.71]; post: 0.95 [-0.16, 2.06]; post+: 0.15 [-0.40, 0.69]; lowest: 0.13 [-1.08, 1.34]) (Niven et al., 2020).
The load-bearing read: it is the intensity, not the intervals
The direction is consistent — HIIE is felt as less pleasant than MICE during, immediately after, at its lowest point, and still (barely) at post+ — yet HIIE and VICE were indistinguishable on affect throughout. Niven draws the inference the decision turns on: «Affective responses to HIIE are similar to responses to VICE during and after exercise, perhaps suggesting that the intensity may explain the responses, rather than the interval or continuous nature of the activity.» (Niven et al., 2020)
The consequence for a recommendation: the aversion is a dose effect of intensity, not a property of the interval format. The interval structure’s recovery periods do not buy back the pleasantness — so do intervals instead of hard continuous work to feel better is not supported, while the discomfort is what high intensity costs, by whatever route is.
No post-exercise rebound to positive — against Dual Mode Theory
Dual Mode Theory predicts that a decline in affect during hard exercise rebounds to positive affect afterwards, which would rescue adherence. This MA finds the opposite: «It was unexpected that at post and post+ exercise the findings indicated a response in favour of MICE, when DMT would suggest that a decline in affect during exercise would rebound to positive affect following activity.» (Niven et al., 2020) The MICE advantage shrinks over time (post+ MD -0.22) but does not flip. The you’ll be glad you did it rebound is not evident against the continuous-moderate alternative.
The enjoyment paradox — real, but discounted for publication bias
Post-exercise enjoyment (a retrospective affective judgement, distinct from in-task valence) ran the other way, favouring HIIE (MD +6.73 [2.45, 11.01]). This is a within-source distinction, not a tension: in-task displeasure and recalled enjoyment are different constructs measured on different scales at different moments, and both can hold. Two cautions keep it from carrying weight on its own: the enjoyment pool was heterogeneous with only 8/15 studies favouring HIIE, and «Inspection of the funnel plot for post-exercise enjoyment (HIIE vs MICE) suggested some asymmetry in favour of positive findings of HIIE» (Niven et al., 2020) — i.e. likely publication bias inflating the pro-HIIE enjoyment signal.
Arousal on its own is not a decision input
HIIE raised arousal (FAS) during and at end of exercise, as expected for harder work. Niven warns against reading this as an affective verdict: «the assessment of arousal on its own is not indicative of an affective response and therefore has limited value. Indeed, it is unsurprising that participants working at a higher intensity report higher arousal.» (Niven et al., 2020) Arousal matters only paired with valence (the circumplex model) — high arousal with unpleasant valence is tension, with pleasant valence is energy. The decision-relevant axis here is valence, which favours MICE.
The gap that blocks the adherence claim
The direction on feeling is settled; the step to behaviour is not. Niven is explicit that the affect-to-future-behaviour link, established for moderate exercise, does not transfer cleanly to HIIE: «A recent study reported that in-task affect during MICE was related to subsequent MICE behaviour in the next 4-weeks, but this relationship was not evident for two HIIE conditions (Stork et al., 2018).» (Niven et al., 2020) So HIIE feels worse, therefore people will do less of it is plausible but unevidenced — the affective substrate is confirmed, the behavioural consequence is a named gap.
Three further limits keep confidence low:
- No moderator analysis for the target stratum. The MA does not separately estimate the sedentary/novice person — the very stratum for whom the inactivity rock is present and adherence is the binding constraint. The pooled effect averages across fitness levels, so the sedentary-specific affect gradient is a type-G gap. A staged companion on the adjacent adherence lever — whether letting people self-select intensity softens the affective cost versus an imposed dose — may bear on this but does not stratify by fitness.
- Acute and lab-bound. Every comparison is a single or short-run bout, mostly cycle-ergometer, in the laboratory — not free-living, repeated, self-selected exercise over months, which is where adherence actually plays out.
- Heterogeneity is large throughout, so each pooled MD is a direction with a wide real spread, not a precise number.
Where this sits in the lever stack
For the sedentary person the big rock is doing any activity at all (Physical Activity Dose and Mortality); modality and intensity are refinements below it. What this page adds at the margin: if the choice is HIIE versus continuous moderate work and the person is intensity-averse, the discomfort is real, intensity-driven, and not rescued by a post-exercise rebound — so prescribing hard intervals to a reluctant novice spends an adherence budget that an easier sustained dose may not. It does not establish that HIIE loses on long-run adherence; that endpoint is unheld. The affective substrate feeds the open adherence question on Measuring and Raising Cardiorespiratory Fitness (Challenge #11) and the adherence caveat on HIIT vs Continuous Training for Type 2 Diabetes.