Ethical considerations

This project involved collecting survey data from pet owners and did not involve experimentation on animals. This study was reviewed by the Institutional Review Board of the University of Pennsylvania on November 21, 2022, and was exempted from Institutional Review Board oversight. The study was approved by Penn Vet’s Privately Owned Animal Protocol Committee.

Study population

Study inclusion criteria were that participating cats were at least 1 year old and had a veterinary diagnosis of chronic inflammatory GI or dermatological disease and had been treated for that disorder within 1 month before study enrollment or had a healthy diagnosis within 3 months before study enrollment. We selected the minimum age requirement of 1 year to ensure that all cats in the study were adults, because temperament is more stable in adult cats.¹⁷

We identified existing Penn Vet patients that had been diagnosed with inflammatory GI disorders (inflammatory bowel disease, chronic enteropathy, and colitis) and dermatological disorders (atopic dermatitis, allergic dermatitis, atopic skin syndrome, chronic pruritus, eosinophilic granuloma complex, pemphigus foliaceus, and chronic otitis externa). To build a robust control group, we used additional data recorded from healthy cats who had participated in a previous study at Penn Vet that had similar parameters (ie, cat owners were asked to complete the Feline Behavioral Assessment & Research Questionnaire [Fe-BARQ] within 3 months of receiving a healthy diagnosis).

We chose to focus on GI and dermatological disorders because previous research has shown that the gut microbiota influences behavior via an immune-mediated response, and it has been suggested that the skin microbiota may also influence behavior.^18–21 In addition, there is some clinical evidence that chronic GI or skin inflammation is associated with behavior changes in other species.^1,22 Finally, GI and skin diseases are present in a substantial percentage of cats treated in the Penn Vet behavior clinic.

Data collection

We assessed the cats’ behavior in their home environments by asking pet owners to complete the Fe-BARQ,²³ a validated behavioral assessment survey, which has been shown to provide information on the temperament and behavior history of cats.²⁴ The survey is available online, easily accessible for users, and takes 10 to 15 minutes to complete. Participating owners were required to be 18 years of 9 age or older. The data collection period was March 15, 2023, to May 31, 2023.

The Fe-BARQ survey includes 100 questions that ask cat owners to rate how often their cat performs a specific behavior. The owners rank the frequency of each behavior on a 5-point ordinal scale (never = 0, seldom = 1, sometimes = 2, usually = 3, and always = 4). These 100 questions measure 38 categories of cat activity, behavior, and interactions with humans and other animals at home. We selected 25 of the behavior categories to compare between groups: 17 subscales (activity, sociability to people, vocalizations, purring, attention-seeking behavior, aggression to strangers, aggression to owner, fear of novelty, separation problem behaviors, trainability, prey interest, location preferences, compulsive grooming, other compulsive behaviors, inappropriate elimination, elimination preferences, and crepuscular activity) and 8 miscellaneous items (agitation, loud vocalizations, spraying, picked up/held, adapts to changes, loud noise response, scratching on inappropriate objects, and kneading behavior). The categories we chose represent a range of behaviors that paint a picture of the cats’ affective state, reactivity, expression of innate feline behaviors, and attention/care-soliciting behaviors.

Data analysis

We first examined the distribution of the data using visual examination of histograms. As many variables were not normally distributed, we used SPSS Statistics software (IBM Corp) and performed the Mann-Whitney U test to compare the Fe-BARQ survey results of cats with inflammatory GI or skin conditions to those of healthy controls.

Several cats in the inflammatory group were being treated with medications that have known psychoactive effects (gabapentin and mirtazapine). Mirtazapine (a tetracyclic antidepressant) is commonly prescribed as an appetite stimulant and antiemetic in cats, and it may also have a calming or sedating effect, while gabapentin (an anticonvulsant) has multiple effects and is commonly used to treat seizures, anxiety, pain, and pruritus in cats.^14,25 Because these drugs may affect the mood and behavior of cats in our study, we analyzed the data twice for each comparison between groups: first with all cats and then we removed the cats receiving psychoactive medications and ran the analyses again.

We used a Kruskal-Wallis test to compare Fe-BARQ scores between cats with inflammatory disease who were treated with corticosteroids, cats with inflammatory disease who were not treated with corticosteroids, and healthy cats, and we used a Dunn test for post hoc analyses. We considered a value of P < .05 to be significant.

Group designation

To begin the analysis, it was necessary to group our subjects by inflammatory status. First, we categorized the cats into 2 groups: healthy (n = 58) and inflammatory (39). Subsequently, we divided inflammatory subjects according to the source of the inflammation: dermatology (n = 17) and GI (22). Twenty-six cats in the inflammatory group had comorbidities that were distinct from inflammatory GI or skin disease. We compared the Fe-BARQ scores of cats with and without comorbidities to explore whether the presence of comorbidities affected the results of this study. We saw no difference between cats with and without comorbidities, and therefore, we did not report that analysis in our results.

We further categorized inflammatory subjects based on pharmaceutical treatments. Eighteen cats in the inflammatory group (46%) were being treated with medications that could potentially affect behavior, including corticosteroids (n = 12) and psychoactive drugs such as anti-anxiety/seizure medications and antidepressants (9).¹⁴ Cats categorized into the steroid group were, at the time of data collection, consistently receiving corticosteroid medications (prednisolone or dexamethasone) daily (n = 10) or every second day (2).

No cats in the healthy group were being treated with medications. Thus, we divided subjects into new groups: inflammatory without psychoactive drugs (n = 30), dermatology without psychoactive drugs (14), GI without psychoactive drugs (16), inflammatory steroid (12), inflammatory nonsteroid (27), and inflammatory nonsteroid without psychoactive drugs (Supplementary Figure S1). Only 1 cat in the steroid group was also receiving a psychoactive drug, and therefore, we did not run the analysis of the steroid group without psychoactive drugs.

Comparison of behavior in healthy cats and those with inflammatory disease

The first comparison between the healthy group (n = 58) and the inflammatory group (39) revealed 3 statistically significant differences. The inflammatory group displayed higher mean ranks in the Fe-BARQ categories of purring (U = 1,396.50, Z = 2.03, P = .042) with a small effect size (r = 0.21), trainability (U = 1,303.50, Z = 2.33, P = .021) with a medium effect size (r = 0.24), and compulsive grooming (U = 1,736.00, Z = 4.91, P < .001) with a large effect size (r = 0. 50), while the remaining 22 categories showed no statistically significant difference between groups (Table 1).

The Fe-BARQ category of purring asks pet owners to report the likelihood of their cat to purr when stroked or petted and when sitting/lying on a person’s lap. The category of trainability asks owners to report how well the cat (1) comes when called; (2) readily responds to simple verbal commands (out, in, quiet, down, up, no, lie down, etc); and (3) attends and listens closely to everything [the owners] say or do. The compulsive grooming category asks owners to report how often a cat exhibits (1) excessive and intensive grooming (eg, grooming several times a day for long periods); (2) self-mutilation, hair barbering (pulls fur with teeth, vigorously nibbles, and/or licks its body parts); and/or (3) sudden frantic licking or chewing²³

When we removed cats treated with psychoactive drugs from the inflammatory group, the same 3 categories showed statistically significant differences between groups: purring (U = 1,208.50, Z = 3.11, P = .002) with a medium effect size (r = .33), trainability (U = 1,030.50, Z = 2.42, P = .015) with a medium effect size (r = 0.26), and compulsive grooming (U = 1,422.00, Z = 5.34, P < .001) with a large effect size (r = 0.57; Table 1).

Comparison of behavior in healthy cats and those with dermatological disease

The dermatology group showed a higher mean rank score for compulsive grooming behavior than the healthy group (U = 910.50, Z = 5.83, P < .001) with a large effect size (r = 0.67). The remaining 24 categories showed no statistically significant differences between groups. This result remained the same when we removed animals treated with psychoactive drugs (Table 1). The categories of trainability (U = 557.00, Z = 1.61, P = .11) and purring (U = 604.00, Z = 1.46, P = .15), both of which had statistically significant differences in the initial inflammatory to healthy group comparison, still showed signs of group difference with small effect sizes (r = 0.19 for trainability and r = 0.17 for purring) but did not reach statistical significance in the comparison of dermatology to healthy groups.

Comparison of behavior in healthy cats and those with GI disease

We observed a similar result in the initial comparison between cats with GI disorders and healthy cats. The GI group showed a higher incidence of compulsive grooming behavior (U = 825.50, Z = 2.37, P = .02) with a medium effect size (r = 0.27) and trainability (U = 746.50, Z = 2.07, P = .038) with a small effect size (r = 0.23) when compared to the healthy group, although no other categories revealed differences between groups (Table 1).

We saw the most notable effect of psychoactive drugs in the comparison of GI to healthy groups. Interestingly, when we removed the cats treated with psychoactive drugs, the GI group scored significantly higher than the healthy group in 4 behavior categories. Compulsive grooming remained statistically significant, while trainability no longer reached statistical significance (P = .052). Three additional categories showed significant group differences: sociability to people (U = 367.50, Z = 2.21, P = .025) with a medium effect size (r = 0.26), purring (U = 686.00, Z = 3.04, P = .002) with a medium effect size (r = 0.35), and fear of novelty (U = 603.00, Z = 2.14, P = .032) with a medium effect size (r = 0.25; Table 1; Figure 1).

View Full Size

Significant behavioral measures between healthy cats and cats with gastrointestinal disease, with and without psychoactive drugs. Cats with gastrointestinal (GI) disease showed significantly higher Feline Behavioral Assessment & Research Questionnaire (Fe-BARQ) score mean ranks than healthy cats in the categories of compulsive grooming and trainability. When we removed cats receiving psychoactive drugs, trainability no longer reached statistical significance, and 3 additional categories showed significance (sociability, purring, and fear of novelty). Statistics performed by Mann-Whitney U test comparing cats with GI disease (with and without psychoactive drugs [PDs]) to healthy cats for each Fe-BARQ category.

Citation: American Journal of Veterinary Research 86, 3; 10.2460/ajvr.24.10.0312

• Download Figure
• Download figure as PowerPoint slide

The sociability to people category in the Fe-BARQ asks owners to report how likely the cat is to be comfortable and relaxed among people in social gatherings, when being petted by unfamiliar (nonhousehold) person(s), and when playing with familiar or unfamiliar children or adults and how likely the cat is to greet unfamiliar adults or children visiting their home in a friendly manner (sniffs, rubs, purrs, and lies on the floor). The Fe-BARQ classifies fear of novelty as the cat displaying restlessness (active investigation) when its resting area is modified (eg, objects moved from usual place and changing fabrics/sheets, etc), as well as restlessness and hypervigilance when unfamiliar objects are introduced into the home.²³

Comparison of behavior in healthy cats, inflammatory cats treated with corticosteroids, and those treated without corticosteroids

To establish whether steroid treatment affected the differences, we observed in the initial comparison between the healthy group and the inflammatory group, and we compared the scores from the steroid group (n = 12), the nonsteroid group (27), and the healthy group (58). Not all of the healthy group participants responded to all Fe-BARQ questions, with 42 participants in the healthy group answering questions related to separation behavior and 57 participants answering questions related to compulsive grooming. Questions that were unanswered were not included in the data analysis.

The analysis revealed a statistically significant difference in separation problem behaviors (χ²[2] = 8.22, P = .016) with a medium effect size (η² = 0.08), with post hoc analysis showing the steroid group scored significantly higher than both the nonsteroid (P = .005) and healthy groups (P = .014). There was no difference in separation behaviors between the nonsteroid and healthy groups (P = .51).

The separation behavior category of the Fe-BARQ asks owners to report how likely the cat is to become restless, agitated and/or pace, sulk, hide, or slink away when the owner prepares to leave the home and if the cat lies down or stays still in the vicinity of the entrance door when the owner prepares to leave. It also records if the cat displays restlessness (active investigation), remains still and alert or hypervigilant (constant ear movements and watchful eyes), and vocalizes by crying or meowing when it is left alone at home.²³

Compulsive grooming also differed between the groups (χ²[2] = 25.35, P < .001) with a large effect size (η² = 0.25), with both inflammatory groups scoring higher than the healthy group (P < .001), but there was no difference between the steroid and nonsteroid groups (P = .26; Figure 2). Unlike in the first analysis where we compared all cats with inflammatory disease to healthy cats, here the difference in trainability did not reach statistical significance (χ²[2] = 5.42, P = .066).

View Full Size

Significant behavioral measures between healthy cats, inflammatory cats treated without steroids, and inflammatory cats receiving corticosteroid treatment. Inflammatory disease cats treated with corticosteroids showed significantly higher Fe-BARQ score mean ranks than healthy cats and inflammatory cats not treated with steroids in the categories of separation behaviors and compulsive grooming. Statistics performed by the Kruskal-Wallis test comparing healthy, steroid, and nonsteroid groups for each Fe-BARQ category.

Citation: American Journal of Veterinary Research 86, 3; 10.2460/ajvr.24.10.0312

• Download Figure
• Download figure as PowerPoint slide

When we removed the cats treated with psychoactive drugs, we found similar differences in separation-related problem behaviors (χ²[2] = 9.93, P = .007) with a medium effect size (η² = 0.12) and compulsive grooming (χ²[2] = 29.27, P < .001) with a large effect size (η² = 0.33). There was also a significant difference in purring (χ²[2] = 10.66, P = .005; Figure 2) with a medium effect size (η² = 0.10). Post hoc tests showed purring was significantly higher among the nonsteroid inflammatory group compared with the healthy group (P = .002), but there were no differences between the steroid and healthy groups (P = .163) or the steroid and nonsteroid groups (P = .318). Trainability, again, failed to reach the threshold for statistical significance (χ²[2] = 5.92, P = .052).

Attention-seeking and care-soliciting behaviors

The higher incidence of purring in the inflammatory group appears to relate to the cats’ attempt to solicit comfort from their owners or to self-soothe, and thus, we interpret this result to indicate a state of increased discomfort and/or stress in the inflammatory group. Although people often consider purring to be a sign of contentment, previous research²⁶ has shown that a primary function of purring is to solicit comfort and attention from others. Additionally, purring may produce a comforting or healing effect on the cat itself, particularly during times of physical and emotional distress.²⁷

If purring in the context of this study were an indication of contentment, then we would expect cats receiving psychoactive medications to purr more. However, we observed the opposite: treatment with psychoactive medications reduced purring in the inflammatory group. Corticosteroid treatment also had a notable effect on purring behavior. The inflammatory nonsteroid group purred significantly more than healthy cats, whereas we observed no group difference between steroid and healthy groups, suggesting that treatment with corticosteroids reduces the physical discomfort associated with the inflammatory response and therefore also reduces the cats’ need to solicit comfort or self-soothe.

The inflammatory and GI groups also scored significantly higher than the healthy group for trainability. One explanation for this difference is an intensified desire for comfort and attention. The Fe-BARQ does not ask owners to report their training methods; however, positive reinforcement works well for training cats and typically involves the owner rewarding the cat with treats, attention, and affection when it responds appropriately to verbal cues.²⁸ Increased trainability suggests that cats with inflammatory disease are more likely to be highly motivated by food, attention, and affection than healthy cats.

Psychoactive drugs did not affect trainability scores in the overall inflammatory group, suggesting that group difference may be related to inflammatory response but not specifically to anxiety or a negative affective state. In the GI compared to healthy analysis, GI scored higher than healthy in trainability after we removed cats receiving psychoactive drugs (P = .052), but the result failed to reach statistical significance. We suspect that the sample size may be hindering the analysis. Steroid treatment, however, reduced trainability scores in the inflammatory group to insignificant levels. Because corticosteroid treatments, which reduce inflammation, also reduced cat trainability scores, this suggests that inflammatory response is indeed associated with trainability.

The GI group without psychoactive drugs scored higher in the category of sociability to people than the healthy group. Previous work has reported differences between levels of sociability in particular cat breeds. For example, Siamese, Abyssinian, and Tonkinese cats are reported to be more sociable toward people, whereas Persian cats are reported to be less playful and sociable toward people.^24,29 It is worth noting that there is some variability within the data from different studies. For example, 1 study²⁴ found Siamese cats to be less social to people. The current study’s population contains a small number of pure-bred cats (n = 7) within the total study population (97; Supplementary Table S1). We therefore consider it to be unlikely that the results for sociability relate to specific breed characteristics.

Sociability to people has also been reported to increase with cat age.³⁰ This study population included a broad range of ages (1 to 15 years) for both the inflammatory and the healthy groups. The average age of cats in both the GI and healthy groups was 8.2 years old. Therefore, we suspect that the observed group difference was unrelated to age.

The result raises a number of questions about the function of feline sociability to people. Gastrointestinal disease has been associated with extreme, insatiable hunger in cats and can result in increased begging for food and therefore more frequent social interactions with people.²⁵ From a physiological standpoint, the increase in sociability to people could relate to the cats’ attempt to alleviate GI discomfort by soliciting food.

From a psychological perspective, increased social interactions are typically associated with cats experiencing neutral or positive emotional states,^31–33 whereas cats experiencing stress or fear are typically less likely to interact.³⁴ Thus, we would expect that mitigating anxious and depressive symptoms through psychoactive therapies would increase sociability. However, the presence of psychoactive drugs in the GI group reduced sociability to people to insignificant levels. This is consistent with the authors’ experience in behavior clinics, where it is frequently reported that psychoactive drugs affect a cat’s ability to communicate.

It could be that interacting with people relieves stress/anxiety in cats and that the function of increased sociability in the GI group is etepimeletic. Increased interactions with people, including affectionate behaviors, such as vocalizing and allorubbing, have been documented in cats experiencing a range of painful and uncomfortable conditions, such as osteoarthritis, hyperthyroidism, and kidney disease.^25,30 Thus, these behaviors may appear to be affiliative and sociable but may in fact be displacement or coping behaviors. Alternatively, as both gabapentin and mirtazapine can have a sedative effect on cats,^14,25,35 it is possible that cats receiving these medications engage in fewer social interactions with people because they are feeling sleepy, sedated, or indeed numb to all stimuli.

Overall, the observed increased rate of purring, trainability, and sociability to people in the inflammatory group reveals that cats experiencing a chronic inflammatory response display more comfort-soliciting behaviors than healthy cats.

Behavioral indicators of anxiety

We observed the most consistent behavioral difference between groups in the category of compulsive grooming, which remained statistically significant across all inflammatory to healthy group analyses (inflammatory to healthy, dermatology to healthy, GI to healthy, steroid to healthy, and nonsteroid to healthy). We found the largest group difference between the dermatology and healthy groups, with the dermatology group being approximately twice as likely to engage in compulsive grooming as healthy cats (Table 1).

Excessive or compulsive grooming in cats may stem from an underlying medical disorder or it may be a displacement behavior resulting from stress or anxiety.^36–39 In this study, all cats in the inflammatory group have known medical conditions that are likely to cause physical discomfort. If physical discomfort were the sole cause of compulsive grooming, then corticosteroid treatment, which mitigates inflammatory response and the associated pain/itching, would reduce compulsive grooming. Similarly, if the behavior were purely stress related, we would expect psychoactive drugs to reduce the incidence of compulsive grooming. However, we observed that across the board, inflammatory disease cats compulsively groomed more than healthy cats, regardless of whether they were receiving corticosteroids or psychoactive medications. We suspect inflammatory cats experience a combination of physical discomfort and a higher level of stress than the healthy group, but more work is needed to determine the specific function of this behavior.

The presence of psychoactive drugs in the GI group reduced fear of novelty insignificant levels, but cats not treated with psychoactive drugs were significantly more fearful of novelty than healthy cats. This finding is consistent with previous work⁷ in dogs that found that dogs with inflammatory bowel disease had increased fear of novel stimuli.

In previous research, the novel object test has been shown as a valid method for assessing fear in several animal species. The test has been performed by other researchers who exposed animals to an unknown object and recorded their behavior to assess the relative degree of fear and curiosity the animal experiences.⁴⁰ Less fearful animals are reported to be more willing to explore and interact with novel objects.⁴¹

In the current study, cat owners answered Fe-BARQ questions that evaluated how often their cat displays restlessness (active investigation) when its resting area is modified (eg, objects moved from usual place, changing fabrics/sheets, etc), as well as restlessness and hypervigilance when unfamiliar objects are introduced into the home.²³

The GI cats’ reactivity to novel objects raises important questions about their overall affective state and potential predisposition toward negative judgement biases. Individuals experiencing a negative emotional state are more likely to interpret a neutral stimulus as negative than individuals experiencing a positive emotional state.⁴² Thus, when cats with GI disease react more fearfully to novel objects (neutral stimulus) than healthy cats, we suggest they are experiencing negative affective states (greater stress or anxiety).

The increased rate of compulsive grooming and fear of novelty support the interpretation that anxiety is present at a higher rate in the inflammatory group than in the healthy group.

Effects of corticosteroid treatment on the behavior of cats with inflammatory disease

The second aim of this project was to investigate behavioral effects of corticosteroid treatment in cats. Corticosteroids have potent anti-inflammatory properties and are reported to be one of the most frequently prescribed drug therapies in veterinary practice.¹³ However, the existing literature describes variability in the effectiveness of corticosteroid treatment in cats. One study⁴³ of feline atopic dermatitis reported that systemic glucocorticoid treatments showed a good clinical response in only 55% of cats. This effect may relate to the bioavailability and potency of the specific drugs used, which were not named in the study. For example, studies that compared the efficacy of specific corticosteroids in cats found that prednisolone has a significantly higher bioavailability than prednisone⁴⁴ and triamcinolone is approximately 7 times more potent than methylprednisolone.⁴⁵

In the current study, all cats in the steroid treatment group received regular (daily or every second day) treatments of prednisolone (n = 9) or dexamethasone (3). The steroid group showed more separation problem behaviors than both the nonsteroid and healthy groups. This remained consistent when we removed cats treated with psychoactive drugs. Separation problems are recognized indicators of fear and anxiety in cats.^46–48 Increased separation behaviors in the steroid group suggests that steroid treatment in cats correlates with increased fear and/or anxiety. Because these behaviors are only present in the steroid group, this suggests a positive correlation between corticosteroid treatment and increased anxiety in cats.

Based on the existing literature, it is possible that the efficacy of steroid treatment might vary between individuals within the steroid group. Both steroid and nonsteroid treatment groups scored higher than healthy cats in compulsive grooming. Therefore, the excessive grooming appears to relate more to inflammatory response than steroid treatment. Similarly, when we removed psychoactive drugs from the analysis, both the steroid and nonsteroid groups showed higher rates of purring than the healthy group. This increase in etepimeletic behavior appears to relate more to inflammatory response than steroid treatment specifically.

Limitations

A limitation of this study is the relatively small sample size for the inflammatory group (n = 39) and healthy group (58), and group sizes decreased further as the inflammatory population was further subdivided based on disease type and medications. Therefore, we acknowledge that the results of our comparisons between groups for each category may not be representative of all inflammatory and healthy populations. Additionally, given that this is a survey-based study, the results are based on owners reporting their perception of their cats’ activities and behaviors. This subjective self-reporting leads to potential unconscious inaccuracies. Nonetheless, owners who continuously interact with and observe their cats in their home environments are uniquely positioned to provide valuable information about cat behavior that cannot be collected in a clinical environment. In addition, the authors acknowledge that a broad range of age-related behavior differences may exist within the study population, which included cats ranging in age from young adults to geriatrics, although we tried to build a study population with similar age variability between groups (average age the healthy group was 8.2 years and inflammatory group was 8.3 years), and therefore, we suspect that the observed group differences were unrelated to age. Further limitations include comorbidities and medications. Many of the cats in the inflammatory group had comorbidities that were distinct from inflammatory GI disease or skin disease. Although we saw no measurable differences between cats with comorbidities and those without, the presence of other diseases may have an effect on behavior. There were also many cats in the inflammatory group being treated with medications other than the corticosteroids and psychoactive medications considered relevant to this study. Future work should examine our results further with larger sample sizes for all groups, attempt to adjust for comorbidities and other medication effects, and attempt to use more objective measures to interpret the motivation for these behaviors.

In conclusion, this study revealed that cats with inflammatory GI or skin diseases display behavior profiles that are different from those of healthy cats. We found a positive correlation between inflammatory GI or skin disease and anxious/comfort-soliciting behavior in cats. Additionally, there was a higher incidence of separation problem behaviors in corticosteroid-treated cats, suggesting that corticosteroid treatment may increase anxiety in cats. These findings have important implications for clinicians and cat owners, as they suggest that monitoring anxious and comfort-soliciting behaviors may be relevant when assessing inflammatory disease in cats. The findings of this study do not rule out the possibility of a medical cause for these behaviors; however, we cannot assume that the observed behavioral differences result solely from an attempt to alleviate physical discomfort. It is beyond the scope of the study to conclusively determine whether these behaviors are a reaction to physical discomfort, emotional distress, or a combination of both. This study demonstrates that additional studies are needed to further explore the effect of GI disease, dermatological disease, and corticosteroid treatments on cat behavior.