The effect of low-intensity exercise on emotional and cognitive engagement in the classroom

The study addresses whether brief, low-intensity physical exercise embedded within university lectures enhances students’ classroom motivation, focusing on emotional and cognitive engagement. Prior research shows benefits of physical activity for mental health but mixed evidence for effects on academic performance, suggesting a need to probe mechanisms such as motivation. The authors hypothesized that a 3-minute, low-intensity ‘slow aerobics’ exercise performed during class would improve emotional engagement (vigour, fatigue, depression, and interest) and cognitive engagement (reduced mind-wandering and sleepiness), potentially via physiological arousal pathways (e.g., norepinephrine) that influence mood and attention.

Evidence indicates physical activity benefits physical and mental health, including reduced stress, anxiety, and depression and improved mood (e.g., increased vigour, decreased fatigue). Mechanistic accounts emphasize exercise-induced changes in monoamine neurotransmitters (notably norepinephrine and serotonin) that modulate arousal and mood. Meta-analytic findings on exercise and academic performance are mixed: some show small positive effects while others find no overall effect, underscoring the need to examine mediators like motivation. Exercise can influence cognitive functions such as attention, likely through locus coeruleus–norepinephrine pathways that enhance alertness and task-focused attention. In educational contexts, mind-wandering is frequent and linked to poorer learning; daytime sleepiness is prevalent among university students and relates to lower academic performance and motivation. While moderate-to-high intensity exercise has been associated with cognitive benefits, evidence for low-intensity exercise indicates potential improvements in arousal, mood, and some cognitive functions, especially when delivered with music and in groups to enhance adherence and engagement.

Design: Within-subjects, repeated-measures field study over nine lessons (one per week) within a semester-long psychology class. Four lessons were assigned to an exercise condition and five to a control condition; the order was pseudo-randomized and fixed across participants (control, exercise, control, exercise, control, exercise, control, exercise, control). About 20 minutes after class began, a 3-minute intervention occurred, after which the lecture resumed. Class length was 90 minutes. Participants: Initially 149 students consented (private university in Japan, 2018). After exclusions (inconsistent IDs, attended only once, absences/late arrivals coded missing), the analyzed sample comprised 114 undergraduates (men = 31, women = 80, unknown = 3; mean age = 20.46, SD = 1.02), yielding 445 observations in the exercise condition and 565 in the control condition. Ethical approval: Doshisha University Faculty of Psychology Research Ethics Committee. Intervention: Exercise condition involved 3-minute low-intensity ‘slow aerobics’ (group exercise with music) via expert-created DVD, combining upper and lower limb and trunk movements designed to keep average heart rate ≤110 bpm. Movements were practiced in the first two exercise sessions and combined in the last two. Control condition involved watching a 3-minute underwater scene video (e.g., dolphins/whales); the same background music was used in both conditions. Measures: Self-reported emotional engagement (vigour, fatigue, depression, and interest) and cognitive engagement (sleep frequency during class and mind-wandering) were collected at the end of each lesson. Scales and ranges: vigour, fatigue, depression (1–5); interest (1–7); sleep frequency (0–6); mind-wandering (1–10). Each construct was assessed by up to three items to minimize participant burden. Statistical analysis: Linear mixed-effects models predicted each dependent variable from condition (exercise = 1, control = −1; fixed effect) and linear time (weeks 0–8; mean-centered), with random intercepts for participants and random participant slopes for condition and time. Random intercepts for lessons were tested; only sleep frequency warranted a lesson-level random intercept (χ² = 6.80, df = 1, P < 0.01). Non-normality was tolerated given mixed-model robustness. Convergence issues were addressed by removing covariances among random effects and, if necessary, simplifying random structures while retaining reliable estimates for the condition effect.

• Exercise increased vigour and decreased fatigue during class relative to control. Mixed-effects models showed a significant positive effect of condition on vigour (B ≈ 0.10, t = 3.85, p < 0.001) and a significant negative effect on fatigue (B ≈ −0.09, t = −2.61, p = 0.01).
• No significant effects of exercise were found on depression or interest (ps ≥ 0.08).
• No significant effects of exercise were observed on cognitive engagement measures (sleep frequency, mind-wandering).
• A significant time trend indicated decreasing sleep frequency across weeks (t ≈ −4.82, p < 0.001), with lesson-level variance accounted for in that model.
• Within-person correlations showed that higher vigour was associated with higher interest and lower fatigue, depression, sleep frequency, and mind-wandering; fatigue positively correlated with sleep frequency and negatively with interest; depression positively correlated with mind-wandering and negatively with interest; interest negatively correlated with sleep frequency and mind-wandering; and sleep frequency positively correlated with mind-wandering.

The findings support the hypothesis that a brief, low-intensity exercise embedded in class can enhance emotional aspects of academic motivation, evidenced by increased vigour and reduced fatigue during lectures. This extends laboratory-based evidence on exercise-induced mood benefits to a real classroom context using a highly practical 3-minute intervention. Interest did not increase, consistent with theory that interest is driven by task content rather than contextual factors like exercise. Contrary to expectations, exercise did not significantly improve cognitive engagement (sleep frequency and mind-wandering). Possible explanations include: (a) low-intensity exercise may exert stronger effects on mood than on cognitive control, with many cognitive benefits in prior literature tied to moderate-to-high intensity activity; (b) substantial individual differences and retrospective self-report could have masked small effects on cognitive engagement. The results suggest that exercise-induced positive mood might facilitate specific academic outcomes, potentially those requiring cognitive flexibility rather than strictly analytic processing, indicating boundary conditions for when exercise can enhance academic performance.

A brief, low-intensity (3-minute) in-class exercise increased students’ vigour and reduced fatigue, demonstrating a feasible strategy to bolster emotional engagement during lectures. Although no significant changes were detected in cognitive engagement (sleepiness, mind-wandering) or interest, the mood benefits highlight a plausible mechanism linking physical activity to academic outcomes. Future research should test downstream effects on different types of academic performance, delineate boundary conditions (e.g., task types), employ larger and more diverse samples across educational contexts, and incorporate physiological measures of arousal to clarify mechanisms.

• No pre-lesson baseline assessments of motivation; thus, changes from pre- to post-intervention within class could not be directly evaluated.
• Generalizability is limited: single convenience sample from one psychology class at one university.
• Potential measurement limitations for cognitive engagement: mind-wandering and sleep frequency were assessed retrospectively and may be subject to bias; large individual differences could obscure small effects.
• No physiological measures of arousal (e.g., plasma markers, pupil diameter, salivary alpha-amylase) were collected to test proposed mechanisms.
• The order of exercise and control sessions was fixed across participants (though lesson-level variance was modeled), which may leave residual confounds at the lesson level.