'When hunger makes everything better looking!': The effect of hunger on the aesthetic appreciation of human bodies, faces and objects

Hunger is a dominant homeostatic drive that shapes perception, action, and cognition. Prior research shows fasting increases attraction to food and acquisition-related behaviors and may influence judgments of physical attractiveness. Most prior studies focused on men’s ratings of women’s bodies, suggesting hunger shifts preference toward heavier female figures, often interpreted via the Insurance Hypothesis (resource scarcity increases preference for adiposity). However, evidence is mixed, with one randomized controlled trial finding no hunger effect and highlighting potential confounds (e.g., alcohol). It is unclear whether hunger-related shifts are specific to body adiposity cues or reflect a domain-general bias toward larger or more bountiful stimuli, and how individual differences (e.g., BMI, perceived hunger) moderate these effects. This study used a within-subject design to test whether fasting versus a controlled snack alters aesthetic liking for slim versus round bodies, faces (as proxies of facial adiposity), and non-food objects (vases), in both female and male observers, and whether perceived hunger and BMI moderate these effects. The authors expected overall higher liking for slim stimuli, but greater liking for rounder stimuli when hungry, potentially specific to bodies/faces (Insurance Hypothesis) or general across categories (bountifulness bias).

Several studies reported that hungry men prefer heavier female body weight compared with satiated men (Nelson & Morrison, 2005; Swami & Tovée, 2006; Pettijohn et al., 2009), aligning with the Insurance Hypothesis, which posits that food insecurity elevates preference for adiposity. However, a randomized controlled trial (Jin et al., 2020) found no hunger effect on male judgments of female attractiveness, implicating potential confounds (e.g., alcohol) in earlier studies and highlighting BMI of raters as a key predictor of attractiveness ratings. Evidence regarding generalization beyond bodies is mixed: Swami et al. (2006) observed no hunger effect on judgments of non-human objects, whereas Saxton et al. (2020) reported that hungrier participants preferred larger bodies and objects, suggesting a possible domain-general bias. Hunger also promotes acquisition of non-food items without necessarily increasing liking (Xu et al., 2015). Individual differences in interoception, cognitive restraint, and adiposity (BMI) may moderate hunger effects (Lattimore, 2005; Stevenson et al., 2015). Prior work also indicates observers’ own BMI shapes attractiveness judgments (Tovée et al., 2000). These mixed findings motivate controlled within-subject manipulations and inclusion of faces and non-body objects to test specificity versus generality of hunger effects and BMI moderation.

Design: Within-subject, two-session experiment manipulating physiological status: fasting (after ≥12 h overnight fast) and snack (immediately after consuming up to two bananas following ≥12 h overnight fast). Session order was randomized and counterbalanced across participants, with sessions held in the morning (08:00–11:30) at least 24 h apart. Participants provided visual analogue scale (VAS) ratings of perceived hunger in each condition; BMI was measured once. Participants: N=44 (21 women), mean age 23.70±0.62 years; mean BMI 23.25±0.49 kg/m². Inclusion: normal/corrected vision, good health; exclusion: psychotropic/vasoactive medication, psychiatric/neurological history, pregnancy, diabetes, allergies to plantain-based foods, specific diet requirements. Mostly right-handed (36/44). Gender identity collected via self-report (female/male). Ethics: Approved by LJMU UREC; written informed consent obtained. Stimuli: Computer-generated 3D images (Poser Pro 2010) of bodies (2 female, 2 male; underwear; face scrambled; frontal/three-quarter views; two postures), faces (2 female, 2 male; hair removed; neutral expressions; frontal/three-quarter views), and vases (7 exemplars; two views). For each category, shape was manipulated to create slim and round versions. Presentation: Three separate blocks (bodies, faces, objects), order counterbalanced and kept constant across sessions. Each block comprised two sub-blocks of 64 trials each (32 slim, 32 round), randomized; total 128 trials per block. Task: Aesthetic liking ratings on a 10-cm VAS (0=not at all; 100=very much); anchor side balanced across participants. Trial timing: fixation 500 ms; stimulus 500 ms (~12°×10° visual angle); mask 500 ms (scrambled version); then rating screen until response. Apparatus: E-Prime 2.0 on Windows PC; viewing distance ~57 cm; 15.6-inch LCD (1024×768, 60 Hz). Hunger manipulation: For the snack session, participants consumed up to two bananas until perceived satiety; VAS hunger ratings collected before/after snack. BMI measurement: height and weight measured with stadiometer and scale; BMI=kg/m². Duration: ~45 min per session. Statistical analysis: Two-tailed t-tests compared self-reported hunger between fasting and snack conditions. Main analyses used 2 (physiological status: fasting, snack) × 3 (stimulus type: body, face, object) × 2 (size: round, slim) repeated-measures ANCOVA on mean VAS liking, with covariates BMI and hunger ratings in fasting and snack conditions. Significant covariate interactions were explored via median splits (e.g., high vs low BMI or hunger) followed by repeated-measures ANOVAs. Bivariate correlations assessed the relationship between BMI and appetite-induced change in aesthetic preference: [(RoundFasting – RoundSnack) – (SlimFasting – SlimSnack)] across all three stimulus types. Alpha=0.05; effect sizes as partial eta squared (ηp²); Duncan post hocs used.

Manipulation check: Self-reported hunger was higher in fasting (M=60.15±25.03) than snack (M=39.94±29.48); t(43)=4.35, p<0.001. Main effects and interactions: - Size: Slim stimuli were preferred over round stimuli; F(1,40)=11.014, p=0.002, ηp²=0.216. - Type × Size: F(2,80)=5.178, p=0.008, ηp²=0.115. Round objects were liked more than round bodies and faces (both ps<0.001; bodies vs faces p=0.771). Slim bodies were liked more than slim faces and objects (both ps<0.001; faces vs objects p=0.287). - Snack hunger covariate × Type: F(2,80)=4.270, p=0.017, ηp²=0.096. Follow-ups (median split on snack hunger, median=30): • Low hunger group (n=19): main effect of type, F(2,36)=6.013, p=0.006, ηp²=0.250. Bodies most liked vs faces and objects (all p<0.025); faces vs objects p=0.305. • High hunger group (n=25): main effect of type, F(2,36)=5.826, p=0.005, ηp²=0.195. Bodies and objects equally liked (p=0.658); faces least liked (all ps<0.004). BMI moderation: - Size × BMI (covariate): F(1,40)=4.127, p=0.049, ηp²=0.094. - Physiological status × Size × BMI: F(1,40)=6.053, p=0.018, ηp²=0.131. Follow-ups (median split on BMI; median=22.63 kg/m²): • High BMI group (n=22; mean 26.02±0.43): Physiological status × Size interaction, F(1,20)=10.647, p=0.004, ηp²=0.347. Slim > round in both fasting and snack (ps<0.001). Critically, round stimuli were liked more during fasting than snack (37.40±1.63 vs 35.07±1.97; p=0.001). Slim stimuli did not differ between fasting and snack (50.94±1.53 vs 51.07±1.92; p=0.683). • Low BMI group (n=22; mean 20.48±0.28): Main effect of size, F(1,22)=87.181, p<0.001, ηp²=0.798; no main effect of physiological status, F(1,22)=0.325, p=0.574; no interaction, F(1,22)=0.050, p=0.825. Correlation: BMI correlated with appetite-induced change in aesthetic preference across all stimuli, r=0.34, p=0.026, indicating higher BMI associated with larger hunger-induced shift toward liking rounder stimuli. Summary: While slim stimuli were generally preferred, fasting increased liking for round stimuli relative to post-snack ratings, not only for bodies but also faces and vases. This hunger-related shift was evident only among higher-BMI participants.

Findings show that physiological hunger modulates aesthetic preferences, replicating and extending prior observations from body attractiveness research. Crucially, the hunger-related shift toward greater liking of roundness generalized across domains: bodies, faces (as proxies of adiposity), and non-corporeal objects. This pattern supports a domain-general bias under hunger toward cues of bountifulness, consistent with accounts that hunger broadens acquisition-related motivations beyond food. At the same time, the moderation by BMI indicates that individual adiposity shapes susceptibility to hunger-induced preference shifts: only higher-BMI individuals exhibited increased liking for round stimuli when fasting. This aligns with aspects of the Insurance Hypothesis (preference for adiposity under perceived scarcity) and with evidence that observers’ BMI influences attractiveness judgments. Potential mechanisms include differences in interoceptive processing and insular cortex engagement associated with BMI, which may alter sensitivity to hunger and satiety cues and affect aesthetic evaluations. The generalization to objects qualifies strictly body-specific interpretations of the Insurance Hypothesis, suggesting hunger may heighten sensitivity to abundance broadly. The results reconcile mixed prior findings by showing that when hunger is experimentally manipulated within subjects and stimulus categories are broadened to faces and objects, domain-general effects emerge, yet are contingent on rater BMI.

This study demonstrates that short-term food depletion increases aesthetic appreciation of roundness across bodies, faces, and objects, indicating a domain-general influence of hunger on liking judgments. Importantly, BMI moderates this effect: higher-BMI observers show hunger-induced shifts toward liking rounder stimuli, whereas lower-BMI observers remain relatively insensitive to physiological status in their ratings. These findings advance understanding of how physiological states and individual adiposity shape aesthetic perception beyond body-specific domains. Future research should incorporate objective biomarkers of hunger/satiety and interoception, examine gender-specific effects and stimulus cues beyond size (e.g., muscularity), assess socio-economic factors, and test stimulus categories in separate sessions to exclude cross-category response biases.

- Gender effects (observer or model) were not analyzed due to sample size constraints; results may differ by gender. - The specific facial and bodily cues driving judgments were not isolated; only size/roundness was manipulated. - Possible perceptual similarity between vases and bodies could contribute to generalization, though vase shapes varied; this cannot be fully excluded. - Socio-economic status/financial security were not assessed and could moderate hunger-related preferences. - All stimulus categories were judged within the same sessions, potentially inducing response biases; although block order was counterbalanced, separate-session designs would better rule out this concern. - Hunger was assessed via self-report VAS rather than physiological biomarkers; interoceptive differences may be underdetected. - Convenience sample of young adults may limit generalizability.