The role of cat eye narrowing movements in cat-human communication

The study investigates whether cat eye-narrowing movements—slow blink sequences—play a role in cat–human communication and whether they signal positive affect. Domestic species benefit from reading and producing human-directed signals. While dogs, horses, pigs, and goats show robust sensitivity to human-given cues and emotional states, cats’ socio-communicative skills remain relatively understudied given their solitary ancestry. Prior work shows cats look to owners in ambiguous situations, can follow human pointing, solicit attention via purring, discriminate their names, and may display facial actions in human-interactive contexts, suggesting potential for interspecific emotional communication. Eye narrowing is a component of positive displays in other species (e.g., canid play, equid and bovine responses to stroking, and the human Duchenne smile), leading to the hypothesis that the cat slow blink sequence may be a positive social signal toward humans. The study tests: (1) whether cats reciprocate human slow blink stimuli with their own slow blink components, and (2) whether slow blink interactions increase cats’ propensity to approach a human.

Research across domestic species shows sensitivity to human-given cues: dogs follow human gestures and attend to attentive states; horses, pigs, and goats can use human cues in object choice tasks and discriminate human emotions, with horses adapting behaviour accordingly. Domestic animals also direct communication to humans (e.g., referential gazing in dogs, horses, and goats). For cats, evidence includes social referencing to owners, performance on human pointing tasks, solicitation purring to manipulate attention, name discrimination even from unfamiliar callers, and facial actions varying with human interaction. Cats may be sensitive to human emotional cues, exhibiting more affiliative allo-rubbing toward depressed owners and adjusting behaviour to human emotional valence. Eye narrowing is implicated in positive affect signalling across species, including canid play/consummatory faces, equids and bovines during stroking or feeding, and the human Duchenne smile, supporting exploration of the slow blink sequence as a positive cat–human signal.

Two experiments were conducted in cats’ home environments. Experiment 1: Subjects: 21 cats from 14 households (10 male, 11 female; age 0.45–16 years, M=7.05, SD=4.59); 14 owners. Inclusion required normal vision and no ocular medical issues. After excluding three dyads as outliers (>2 SD), analyses included 18 cat–owner dyads. Procedure: Owners were trained by experimenters to deliver the slow blink stimulus (series of half-blinks culminating in prolonged eye narrowing or closure). Trials were counterbalanced between (a) slow blink stimulus and (b) no human interaction (owner present but not engaging). Owners sat ~1 m from the cat when delivering the stimulus. Two cameras recorded owner and cat faces (Sony DSC-HX9V for owners; Canon G9 for cats). Trials lasted up to 2 min (mean 62.75 s; range 19.14–120 s). Slow blink delivery averaged 14.58 eye-movement stimuli per minute (range 3–30.6). Control trial mean length: 59.86 s (range 21.03–120). Behavioural coding: CatFACS-based coding for cat half-blink (AU147), blink (AU145), eye closure (AU143), and an added “eye narrowing” code (sustained partial closure for at least 2 frames). Owner eye narrowing and closures were also coded. One certified CatFACS coder (blind to cat condition) coded all videos; a second coder scored 25% blind. Inter-observer reliability Cronbach’s alpha=0.9. Statistics: Because trial lengths varied, rates (count per second) were computed for each eye movement. Linear mixed models in R (lmerTest) compared rates between conditions (fixed factor: Condition; additional fixed factors: Number of Cats in Household, Cat Sex, Cat Age). Random factor: Cat Identity nested within Household. Model selection used AIC; movements occurring in <5 cats were not analyzed. Experiment 2: Subjects: 24 cats recruited; after excluding 6 outliers (>2 SD), 18 cats from 8 households remained (12 male, 12 female originally; age 1–17 years, M=6.00, SD=4.78). Cats were naïve to Experiment 1. Control condition was a neutral facial expression without direct eye contact (modified after pilot trials suggested direct eye contact could be perceived as threatening). Procedure: An unfamiliar experimenter (JF) avoided prior contact with cats. After a 2-min baseline, the experimenter offered a flat hand (palm up) to assess baseline approach tendency, then retracted the hand and either (a) delivered the slow blink stimulus or (b) adopted a neutral face looking slightly away (no direct eye contact). After 1-min trials, the experimenter offered her hand again for a few seconds (mean 3.71 s; no condition difference). Trials were separated by ~2 min intervals; order counterbalanced. Cameras recorded the cat and experimenter (Panasonic HC-X920; Sony DCR-SR37). Behavioural coding: As in Experiment 1, except normal maintenance blinks were omitted from coding. Approach responses were coded as Avoid, Neutral, or Approach. To control for attention calls, cat eye movements within 0.5 s of the experimenter calling the cat’s name (without a concurrent experimenter eye closure) were excluded. Statistics: Trial lengths were fixed (1 min), so raw counts were analyzed via linear mixed models with Condition plus Number of Cats in Household, Sex, and Age as fixed effects; Cat Identity nested within Household as random. Model selection used AIC. Wilcoxon signed-rank tests compared approach tendency across conditions (coded 1=avoid, 2=neutral, 3=approach). Ethical approval was granted by the University of Sussex ERC (Non-ASPA-Nov2013), following ASAB guidelines; informed consent obtained from owners.

Experiment 1: • Eye narrowing rate was higher during the slow blink stimulus than during no human interaction (Experimental M=0.06±0.07; Control M=0.03±0.03; z=−2.64; p=0.017). • Half-blink rate was higher in the slow blink condition than control (Experimental M=0.21±0.15; Control M=0.15±0.06; z=3.45; p=0.003). There was a Sex×Condition interaction: males showed a significant increase in half-blinks in the experimental condition, females showed little effect (Male: Experimental M=0.30±0.16 vs Control M=0.13±0.07; Female: Experimental M=0.13±0.06 vs Control M=0.10±0.05; z=2.51; p=0.02). • Normal blinks showed no condition effect (best-fit null model). • Eye closures were too infrequent (4 cats in experimental; 3 in control) for analysis. Experiment 2: • Half-blink counts were higher during slow blink than neutral face (Experimental M=5.33±6.10; Control M=1.94±1.77; z=2.92; p=0.007). • Eye narrowing counts were higher during slow blink than neutral (Experimental M=2.78±3.73; Control M=0.33±0.97; z=3.35; p=0.002). • Eye closures increased with cat age (z=−2.35; p=0.02). • Cats’ approach scores to the experimenter’s hand were higher after slow blinking than after a neutral face (Slow blink M=2.78±0.43; Neutral M=2.39±0.70; z=2.11; p=0.035).

The findings demonstrate that cats reciprocate human slow blink stimuli with their own eye-narrowing components (half-blinks and sustained narrowing) and that slow blink exchanges are associated with increased approach behaviour toward humans. This pattern supports the hypothesis that slow blink sequences function as positive-affect communication in cat–human interactions. Approach behaviour is commonly used as an index of perceived pleasantness; thus, cats’ greater approach after slow blinking suggests they either experienced a positive affective state during the interaction or evaluated the human more positively afterward, potentially through prior learning. The slow blink sequence’s eye narrowing aligns with positive emotional signals in other species, including the human Duchenne smile and affiliative or pleasurable contexts in canids, horses, and cows, indicating possible cross-species commonalities in positive facial signalling. The results also show that both familiar owners and unfamiliar experimenters can elicit these responses, suggesting generalization across human partners and highlighting the potential utility of slow blinking in promoting positive cat–human interactions.

This study provides the first systematic experimental evidence that cat slow blink sequences constitute a positive social signal in cat–human communication. Cats increased half-blinks and eye narrowing in response to human slow blink stimuli, and were more likely to approach an unfamiliar experimenter following such interactions. These insights can inform strategies to enhance cat welfare and human–cat bonding in homes, shelters, and veterinary contexts. Future research should assess the role of direct gaze, examine slow blinking in cat–cat (conspecific) communication, test mechanisms (learned vs evolved), and perform controlled laboratory replications to isolate contributing cues and distances.

• Experiment 1 involved variability in owners’ execution of the slow blink (differences in delivery, talking, wearing glasses), and unquantified variation in distance/position during control trials, adding noise. • In Experiment 2, the neutral control excluded direct eye contact; differences in gaze direction/attentional state could have influenced approach behaviour, representing a potential confound. • Naturalistic home settings limited precise control over cat positioning and distance, potentially increasing variability. • Some outcomes (eye closures) were too infrequent for analysis in Experiment 1. • Potential learning or reinforcement history with owners could contribute to responses; generalization across humans was only partially addressed with one unfamiliar experimenter. • Outlier exclusion reduced sample sizes (18 dyads in Exp. 1; 18 cats in Exp. 2).