Impact of high-fat diet and feeding time on daily anxious-like behaviors in aged male rats

Abstract

Anxiety symptoms exhibit day-night variation, often peaking in the late afternoon or evening. Despite clinical recognition of anxiety’s diurnal variation, time of day is rarely considered in the design or interpretation of anxiety research. In this study, we used aged (10–12 month) Long-Evans male rats to examine the extent in which time of day, chronic (> 15 weeks) high-fat diet feeding, and time-restricted feeding affect anxious-like behaviors in the elevated plus maze and the novelty box test. We find that anxious-like behaviors, such as closed-arm entries, are consistently higher during the animal’s inactive phase. This day-night variation was unaffected by high-fat diet feeding. Interestingly, restricting food to a 12-hour window did not invert the day-night variation in anxiety but reduced anxious-like behavior overall. These findings underscore the importance of multiple testing times when measuring anxiety and illustrate the relative resiliency of daily patterns in anxiety expression. Overall, we conclude that time of day modulates anxious-like behavior and should be considered in experimental designs and therapeutic interventions for anxiety.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12993-026-00326-0.

Introduction

Clinically, anxiety symptoms exhibit pronounced diurnal variation that peaks in the late afternoon or evening [1,2]. This temporal pattern may reflect an underlying circadian influence [3, 4] and can offer critical insight into anxiety diagnosis and treatment. In rodent models, anxious-like behaviors are frequently tested using the elevated plus maze (EPM) [5–8] where increased time or entries into the closed-arms generally reflects an increase in anxious-like behavior. Conversely, increased time or entries into the open-arms and/or center indicates a decrease in anxious-like behavior. Anxious-like behaviors can also be assessed with a novelty box test (NBT). The NBT is conceptually similar to the open field test in both design and scoring, in that increased time spent in the center and a greater number of center entries are interpreted as reduced anxious-like behavior in rodents. However, the NBT is most notable for its assessment of novelty-seeking and exploratory behaviors [9–11]. Considering diurnal variation when collecting and interpreting anxious-like behavior can increase our mechanistic understanding of anxiety, yet most preclinical studies rarely consider diurnal variation as a variable.

Most physiological molecular and behavioral rhythms are organized in 24-hour patterns called circadian rhythms. Evidence in humans shows that mutations in PER3, a core clock gene involved in generation of molecular circadian rhythms, are associated with an increase in anxiety [12]. This suggests that endogenous rhythms may also affect anxiety presentation in humans. Moreover, external factors known to disrupt circadian biology exacerbate anxious-like behaviors. For example, prolonged light exposure disrupts circadian biology and also increases anxious-like behaviors in male rats [13]. Still, multiple studies fail to report day-night differences in anxious-like behavior [14,15], raising important questions about how and when anxious-like behaviors are assessed. Unlike clinical assessments, many anxious-like behaviors in rodents are measured during the rodent’s inactive (light) phase [16]. If anxious-like behaviors in rodents peak near the onset of rest, as is observed in humans, then testing at a single timepoint may artificially inflate anxiety estimates and obscure meaningful time-of-day effects.

]Obesity is an important, yet often overlooked, factor that affects both anxiety and daily rhythms. Epidemiology evidence suggests that obesity causes a 25% increase in anxiety disorder diagnosis [17]. While this relationship is clinically defined, modeling the relationship between obesity and anxiety in animals requires operational proxies. In rodent studies, obesity is most commonly achieved through high-fat diet feeding and is therefore used as a controlled experimental approach to model obesity-associated physiological and behavioral changes, including anxious-like behaviors. However, it is presently unclear the extent to which high-fat diets affect anxiety and/or the rhythmic expression in anxious-like behaviors. Importantly, high-fat diet feeding is not solely a metabolic manipulation. High-fat diets can disrupt circadian rhythms by altering gene expression [18], locomotor activity [19] and feeding behaviors [20,21] and negatively affect mental health outcomes in both humans [22–24] and rodent models [25,26]. Mechanistically, high-fat feeding increases affective disorders through an increase in neuroinflammatory signaling [25,27], disruption of the gut-brain axis [26,28,29] and dopamine signaling [30]. Given that feeding time is a powerful time cue for circadian biology, aligning feeding time to the active phase through time-restricted feeding may restore circadian regulation and in turn, mitigate anxious-like behaviors associated with high-fat diet induced disruptions.

In the present study, we sought to determine the extent to which day-night variation in anxious-like behavior is affected by high-fat feeding and/or time-restricted feeding. We hypothesized that anxious-like behaviors would exhibit diurnal variation with peak expression occurring during the animal’s inactive phase. To test this, male rats were fed either a standard chow or a 60% kcal from fat high-fat diet and assessed using the EPM and NBT at timepoints corresponding to their active and inactive phase, Data was collected under both ad libitum and time-restricting feeding conditions. We find that anxious-like behaviors display clear time-of-day differences with more anxious-like behaviors (e.g., closed-arm entries) occurring during the inactive phase. Maintaining rats on a high-fat diet did not affect diurnal variation in anxious-like behaviors. Similarly, and in contrast to our expectations, the timing of food presentation during a time-restricted feeding protocol did not organize daily variation in anxious-like behaviors. However, limiting food availability to 12-hours a day, regardless of whether the 12-hour feeding period occurred during the active or inactive phase, reduced anxious-like behaviors. These data demonstrate that the timing of anxiety assessment can meaningfully affect interpretations in rodent models and underscore the importance of incorporating circadian context into the experimental design of anxiety studies. Our results also point to time-restricted feeding as a potential intervention to reduce anxiety.

Methods

Animals. Male, Long-Evans rats (n = 33) were purchased from Charles River at 15 weeks old and tested 10–12 months of age. Upon arrival, all rats were pair-housed and maintained on either chow (21.55% kcal from fat derived from lard, PicoLab 5058) or 60% kcal/fat high-fat diet (60% kcal from fat derived from lard, Teklad, TD.06414). A small subset of rats (n = 2 in the chow-fed group) were single housed to alleviate social dominance fighting. Although small, this represents a potential limitation of the study, as the social housing conditions for these animals differed from those of the group-housed rats. Rats remained in the animal colony on their respective diets for 15 weeks prior to the start of the experiment to ensure obesity development [31]. Rats were housed in 12 L:12D light conditions with lights on at 8 am and off at 8 pm. Zeitgeber Time (ZT) 0 corresponds to lights on and ZT 12 corresponds to lights off. The starting animals in the experiment were a total of 33 but some animals were excluded due to difficulty collecting parameters that were needed for analysis such as blood samples or anyMAZe tracking. All studies were reviewed, approved by, and performed according to the guidelines of the Institutional Animal Care and Use Committee at Marquette University (Milwaukee, WI).

Experimental Design. All animals initially received food ad libitum and were assigned to either a chow or high-fat diet. Animals were then randomized, stratified by body weight, to undergo their first behavioral assessment during either the active or inactive phase. The order of EPM testing during time-restricted feeding (active vs. inactive phase first) was similarly randomized based on body weight. While on ad libitum feeding, the rat’s behavior in the Novelty Box Test (NBT) and the Elevated Plus Maze (EPM) was assessed at two times of day (during the active, ZT 13–14.5, and the inactive phase, ZT 2–3.5). While on ad libitum feeding, but after NBT and EPM testing, a blood sample and a body weight were collected. All animals then had food-restricted to 12-hours a day (see Time-Restricted Feeding). Animals on the high-fat diet were divided into two groups – those fed only during the 12-hour active phase (at night) and those fed only during the 12-hour inactive phase (during the day). Animals on the chow diet were restricted to eating during the 12-hour active phase (at night). Feeding chow-fed rats exclusively during the inactive phase leads to substantial weight loss and torpor (unpublished observations), raising welfare concerns and confounding interpretation. For this reason, feeding during the inactive phase was limited to only high-fat diet-fed animals. After 1 week on the restricted feeding, all animals were reassessed at two times of day in the EPM. Overall, animals were subjected to a total of two NBT exposures and four EPM exposures.

Body weight collection. Each animal was weighed prior to the start of behavioral testing and again immediately before blood collection for a total of two times. The study was not specifically designed to evaluate body weight changes associated with time-restricted feeding. Therefore, body weight measurements were not obtained following time-restricted feeding, which represents a possible limitation in the current study.

Novelty box test (NBT). To assess responses to a novel environment, rats were placed in the novelty box for a 60-minute recording period. The novelty box paradigm closely parallels the open-field test in both apparatus design and behaviors assessment metrics. The novelty box focuses on shorter durations (typically 5–20 min) to capture acute novelty-driven exploration, longer recordings have been used when the goal is to examine behavior beyond initial novelty and habituation. Previous work demonstrates that habituation to a novel environment typically occurs within the first 10–15 min of exposure [32]. Accordingly, extending the recording period allows assessment of sustained behavioral patterns after habituation, which was central to the present study’s focus on chronic anxious-like behavior across circadian phase and dietary conditions. Similar extended recording durations have been used in prior studies assessing novelty-related and open-field–like behaviors [33–42]. Rats (n = 32; chow = 15, HFD = 17) were acclimated to the testing room 2 weeks prior to testing for 2 h each day during the day-time. On the day of the testing, rats were again acclimated to the room for 1 h prior to testing during the active or inactive phase. Testing occurred during the rat’s active (ZT 13–14.5) and inactive phase (ZT 2–3.5) on separate testing days with at least 2 resting days between trials. Testing during the inactive phase (ZT 2–3.5) occurred under white overhead lights (200–250 lx). Testing during the active phase (ZT 13–14.5) occurred under no overhead light (1–10 lx). The dimension of the arena is 16 by 16 by 14 inches. Center time, center distance and total distance were measured and analyzed. A counterbalance design was used such that not all rats began testing in their active phase. For all runs, rats were placed in the center of the novelty box and recorded for an hour [43]. During an animal’s second exposure to the NBT, the arena was cleaned with 70% alcohol solution to remove any remaining scent from the previous animal. Additionally, the pattern on the walls was changed (i.e., to have stripped patterns) to maintain a novel environment. Investigators were not present in the room during the testing. Data were scored using versa max version 2.30-127E and included center time, center distance, and total distance travelled.

Elevated plus maze (EPM). Rats (n = 31; chow = 14, HFD = 17) were acclimated to the testing room 2 weeks prior to testing for 2 h each day during the day-time. On the day of testing, rats were again acclimated to the room for 1 h prior to testing during the active or inactive phase. Testing occurred during the rat’s active (ZT 13–14.5) and inactive phase (ZT 2–3.5) on separate days with at least 3 resting days between trials. Testing during the inactive phase (ZT 2–3.5) occurred under white overhead lights 150–250 lx). Testing during the active phase (ZT 13–14.5) occurred under dim red lights (45–55 lx). A counterbalance design was implemented such that not all rats began testing in their active phase. For all runs, rats were placed at the center of the elevated plus maze and recorded for fifteen minutes. One investigator was present in the room during testing but remained quiet and not visible to the rat. Data were scored using AnyMAZE (Version 7.40) and included closed-arm entries, closed-arm time, open-arm entries, open-arm time, center entries, center time, and distance traveled. Animals with 80%+ of detected frame tracking was verified via visual inspection and included in the analysis. EPM trials with less than 80% frame tracking were excluded from data analysis. The animal was considered to enter an area when 80% or more of the rat was detected entering the area. Considering our study’s goal to characterize anxious-like behavior in the active and inactive phase and reveal any diet-dependent and time-restricted feeding differences, we use a longer (15-minute) recording to account for sustained anxious-like behavior beyond initial novelty to a new environment similar to protocols used previously [44].

Time-restricted feeding. All rats underwent 1 week of restricted feeding before time-restricted feeding data collection began. Restricted access to food was controlled by the investigator through the manual removal and replacement of food. Food access was limited to 12-hours. Water was available all 24 h. HFD “Day-fed” rats (n = 4) were only fed during the rat’s inactive phase (ZT 0 – ZT 12). HFD “Night-fed” rats (n = 11) were only fed during the rat’s active phase (ZT 12 – ZT 24). Rats maintained on the chow diet (n = 11) were only restricted to the “Night-fed” condition to prevent excessive weight loss (unpublished prior observation). Rats on the high-fat diet were split into “Day-fed” and “Night-fed” groups.

Blood collection and analysis. Blood collection occurred 1 week after the first day-night assessments in the elevated plus maze. This timing was selected to capture baseline corticosterone levels associated with chronic dietary exposure rather than an acute stress response to EPM testing. All animals remained in their home cage and were not fasted prior to blood collection. All animals were weighed between ZT 10–11 and then 600 uL total of blood from each animal tail was collected during ZT 11–12 into Microvette tubes and transported on ice and centrifuged (10 min at 3,000 RPM under 4 degree conditions). Plasma was immediately collected and frozen (-80 °C) for later use. Plasma corticosterone and leptin were tested using a rat leptin ELISA kit (Crystal Chem, catalog 90040; sensitivity range is 0.2–12.8 ng/mL) and a rat corticosterone ELISA kit (Crystal Chem, catalog 80554; sensitivity range is 15–2250 ng/mL) according to the manufacturer’s instructions. Animals were excluded from analysis in the event of insufficient plasma. Statistics were preformed using a linear regression with Prism Version 10.

Statistical analysis. All statistics were run using SPSS Version 29.0.1.0 or Prism Version 10. All SPSS statistical values are reported as means ± SEM and were analyzed by Multivariate ANOVA to assess the effects of all variables (Table 1: diet, testing time of day, feeding length, feeding time of day and exposure number) on measured outcomes (closed-arm entries, closed-arm time, center entries, center time, open-arm entries, open-arm time, distance traveled). All Prism statistical analysis values are reported as means ± SEM and were analyzed by unpaired t-test, or linear regression were appropriate. Statistical significance was set at p < 0.05.

Table 1.

Variables used in SPSS statistical analysis

Variable name	Levels	Description
Diet	Standard Chow Diet (chow) High-Fat Diet (HFD)	The diet that rats were fed
Testing Time of Day	active (dark) phase inactive (light) phase	The time that animals were tested in the NBT or EPM
Feeding Length	Ad libitum 12-hours	The amount of time per day that animals had access to food
Feeding Time of Day	Night-fed Day-fed	The rats’ dominant feeding time of day. Day-fed rats are exclusively fed during the inactive phase. Night-fed rats are exclusively fed during the active phase. Ad libitum fed rats predominantly eat during the active phase and are coded as Night-fed.
Exposure Number	1 2 3 4	The number of times the rat was exposed to an individual test (either the NBT or the EPM)

Results

A high-fat diet does not affect the diurnal fluctuation in anxious-like behaviors

Ad libitum fed rats were maintained on either standard chow or 60% kcal from fat high-fat diet (HFD) and behaviors were assessed in the novelty box test (NBT) and elevated plus maze (EPM; Fig. 1). Rats were tested in the NBT and EPM both during their inactive (ZT 2–3.5) and active (ZT 13–14.5) phase (Fig. 1A). When fed ad libitum, diet (regardless of composition) had no main effect on animal behavior in the NBT or the EPM (Feeding Length*Diet; Ad libitum: chow vs. HFD, all p > 0.05; Supplemental Fig. 1: total distance F(1,52) = 1.543, center distance F(1,52) = 0.362, center time F(1,52) = 0.004 and Supplemental Fig. 2: closed-arm time F(1,70) = 0.420, open-arm entries F(1,70) = 0.358, closed-arm entries F(1,70) = 1.332, open-arm time F(1,70) = 0.107, center time F(1,70) = 0.474, center entries F(1,70) = 0.814, total distance traveled F(1,70) = 0.357).

Fig. 1.

Experimental Design. Rats were fed ad libitum for 15 weeks before assessment in the Novelty Box Test (NBT) during both the active (ZT 13-14.5) and inactive (ZT 2-3.5) phase. 3 weeks after the completion of the NBT assessment, rats were assessed in the Elevated Plus Maze (EPM) again during both the active and inactive phase. 1 week after initial testing in the EPM, blood was collected for later analysis. Rats were then divided into three groups and underwent time-restricted feeding: rats on the chow diet were limited to food access only during the active phase (Night-fed), while rats on the high-fat diet were divided into two, one fed only during the active phase (Night-fed) and the other only during the inactive phase (Day-fed). After 1 week of restricted feeding, all animals underwent final EPM testing again during both the active and inactive phase

Time of day had a significant effect on behavior in the EPM but not in the NBT. EPM distance traveled, center entries, and closed-arm entries were greater during the animals’ inactive phase, independent of diet condition (Feeding Length*Testing Time of Day; Ad libitum: inactive vs. active, F(1,70) = 4.032, p = 0.049, F(1,70) = 5.766, p = 0.019, and F(1,70) = 10.026, p = 0.002, respectively; Fig. 2A–C). No significant day-night variations were noted in open-arm entries, (Feeding Length*Diet*Testing Time of Day, all p > 0.05; Chow: F(1,70) = 0.007, HFD: F(1,70) = 0.420, Fig. 2D) closed-arm time, open-arm time, or center time (data not shown). Indeed, rats under both diet conditions exhibited more closed-arm entries during their inactive phase (Feeding Length*Diet*Testing Time of Day; Ad libitum, chow: inactive vs. active, F(1,70) = 5.064, p = 0.028; Ad libitum, HFD: inactive vs. active, F(1,70) 4.967, p = 0.029, Fig. 2C). Together, these behaviors support an increased propensity for anxious-like behavior during the inactive phase of rats. Notably, there was no diurnal variation seen in the behavior recorded from the NBT (Time of Day tested; Active vs. inactive phase, all p > 0.05; total distance F(1,52) = 0.16, center distance F(1,52) = 0.034, center time F(1,52) = 0.404; Supplemental Fig. 3A-C).

Fig. 2.

Anxious-like behaviors vary by time-of-day and are not affected by high-fat feeding. A There is a main effect of distance traveled to be higher during the inactive phase of testing, *p < 0.05. B There is a main effect of center entries to be increased during the inactive phase of testing, *p < 0.05. C Closed-arm entries are increased during the inactive phase of testing as measured both by main effects, *p < 0.05, and within group comparisons, **p < 0.01. D There were no significant differences in open-arm entries, (E) closed-arm time (F) open-arm time (G) or in center time spent. Statistics were performed using SPSS Version 29.0.1.0. All values are reported as means ± SEM and were analyzed by Multivariate ANOVA

Under limited situations, body weight is correlated to anxious-like behaviors

Perhaps due to the age of the rats, body weights taken under ad libitum feeding conditions was highly variable in both diet conditions (Fig. 3A). Therefore, we ran additional analyses to determine the extent to which body weight, as opposed to diet, correlated to behavior in the EPM. We found that body weight was positively correlated to the number of closed-arm entries during the active phase, but not inactive phase, of chow-fed rats (R² = 0.4722, p = 0.0338, Fig. 3B). This relationship was not observed in high-fat fed rats (R² = 0.0786, p = 0.3774, Fig. 3B). No other behavior recorded during the EPM correlated to body weight (data not shown). Next, we wanted to understand how a high-fat diet consumption affects corticosterone and leptin levels. We found that plasma levels of corticosterone (R² = 0.2576, p = 0.0096; Fig. 3C) and leptin (R² = 0.3343, p = 0.0010; Fig. 3D) were highly correlated to body weight.

Fig. 3.

Body weight is positively correlated with an increase in anxious-like behaviors in chow animals. A At the time of testing, high-fat ad libitum fed animals weighed more than their chow fed counterparts (p = 0.0006; R² = 0.3562). B Only chow-fed rats displayed a body weight-dependent relationship with anxious-like behavior, as indicated by closed-arm entries during the active phase (R² = 0.4722, p = 0.0338). No relationship was observed in high-fat fed rats (R² = 0.0786, p = 0.3774). C Corticosterone levels negatively correlate with body weight (R² = 0.2576, p = 0.0096). D Leptin levels positively correlate with body weight (R² = 0.3343, p = 0.0010). Statistics were performed in Prism Version 10. All values are reported as means ± SEM and were analyzed by unpaired t-test, or linear regression were appropriate

Limiting food availability to 12-hours a day is associated with a decrease in anxious-like behavior

Time-restricted feeding, a protocol that limits food availability to 12-hours a day without necessarily restricting caloric intake, has numerous health benefits in both humans [45,46] and rodents [47–49]. The effect of time-restricted feeding on anxious-like behaviors has yielded mixed results [50–52]. To determine how time-restricted feeding affects anxious-like behaviors, all rats had food availability reduced to 12-hours a day while remaining on their respective diets of either HFD or chow. After 1 week of exclusive day-feeding or night-feeding (Fig. 1) behavior in the EPM was reassessed during the active and inactive phases.

Overall, restricting food to a 12-hour window is associated with a decrease in anxious-like behaviors largely independent of diet. There was a main effect of food restriction, regardless of the food timing, to reduce time spent in closed-arms and increase in time spent in the center independent of diet (Feeding Length; Ad libitum vs. 12-hours, F(1,70) = 6.868, p = 0.011 and F(1,70) = 13.755, p < 0.001, respectively, Fig. 4A, B). Restricted feeding decreased the time spent in the closed-arms for high-fat diet-fed animals and trended to decrease the time spent in closed-arms for animals maintained on the chow diet (Diet*Feeding length; Chow: Ad Libitum vs. 12-hour feeding, F(1,70) = 2.046, p = 0.050; HFD: Ad Libitum vs. 12-hour feeding, F(1,70) = 3.932, p = 0.033; Fig. 4A). Both chow and HFD-fed rats spent more time in the center when food was restricted to 12-hours (Diet*Feeding length; Chow: Ad libitum vs. 12-hour feeding, F(1,70) = 4.33, p = 0.041; HFD: Ad Libitum vs. 12-hour feeding, F(1,70) = 8.013, p = 0.006; Fig. 4B). Since the above comparisons included both Night-fed and Day-fed conditions, and therefore introduce circadian disruption as a confounding variable, we next examined the specific effects of Night-feeding using a within-subject comparison of EPM behavior in Night-fed animals relative to their prior ad libitum data. Night-feeding increased time spent in the open-arms and decreased time spent in closed-arms (Feeding time*Feeding Length; Night-fed: Ad libitum vs. 12-hours, F(1,70) = 5.555 p = 0.017 and F(1,70) = 4.775p = 0.01, respectively; Fig. 4C, D). Taken together, our data suggest that restricting feeding time to a 12-hour window during the animal’s active phase reduces anxious-like behaviors in rats fed both chow and high-fat diets. Due to the rats’ continued exposure to the elevated plus maze, considering the potential contribution of habituation in the time-restricted feeding sector is necessary since it cannot be entirely ruled out.

Fig. 4.

Time-restricted feeding reduces anxiety-like behaviors, with night-feeding driving the effect. A There is a main effect of time-restricted feeding to decrease time spend in the closed-arms, *p < 0.05 with a significant within group comparison for animals maintained on a HFD; * p < 0.05. B Center time spent is increased during time-restricted feeding as measured by both main effects, ***p < 0.001, and within group comparisons, *p < 0.05 and **p < 0.01 respectively. C Exclusive Night-fed conditions increase open arm time spent compared to ad libitum; *p < 0.05. D Exclusive night-fed conditions decrease closed-arm time spent compared to ad libitum; **p < 0.01. Statistics were performed using SPSS Version 29.0.1.0 All values are reported as means ± SEM and were analyzed by Multivariate ANOVA

Day-night differences in EPM behavior were also observed during time-restricted feeding. Regardless of diet, restricting food to a 12-hour window increased time spent in the center during both active and inactive phases (Testing time of Day* Feeding Length; AM: Ad Libitum vs. 12-hour feeding, F(1,70) = 6.060, p = 0.016; PM: Ad Libitum vs. 12-hour feeding, F(1,70) = 7.760, p = 0.007; Fig. 5A). Time-restricted feeding also decreased time spent in the closed-arms but during the animal’s active phase (Testing time of Day* Feeding Length; PM: Ad Libitum vs. 12-hour feeding, F(1,70) = 4.404, p = 0.039; Fig. 5B). Moreover, time-restricting feeding affected day-night variation in EPM behavior in a diet-dependent manner. Rats fed a high-fat diet maintained diurnal variation in center entries, closed-arm entries, and distance traveled (Feeding Length*Diet*Testing time of Day; 12-hour fed, HFD: active vs. inactive phase, F(1,70) = 14.492 and p < 0.001, F(1,70) = 17.732, p < 0.001, F(1,70) = 8.058, p = 0.006, respectively, Fig. 5C, E) similar to their behavior during ad libitum feeding. However, chow-fed rats lost these diurnal variations when restricted to a 12-hour window of feeding (Feeding Length*Diet*Testing time of Day; 12-hour fed, Chow: active vs. inactive phase, center entries F(1,70) = 0.360, p > 0.05, closed-arm entries F(1,70) = 0.809, p > 0.05, and distance traveled F(1,70) = 0.254, p > 0.05, Fig. 5C-E).

Fig. 5.

Time-restricted feeding decreases anxious-like behaviors during the active phase while selectively influencing animals on a HFD. A Restricted feeding increases center time spent during both the active and inactive phases; *p < 0.05 and **p < 0.01. B Restricted feeding decreases closed-arm time only during the active phase, *p < 0.05. C While on time-restricted feeding, there is a main effect of testing time of day with more center entries occurring during the inactive phase; *p < 0.05 and a significant within group comparison for animals maintained on the HFD; ***p < 0.001. D While on time-restricted feeding, there is a main effect in testing time of day with more closed-arm entries occurring during the inactive phase; **p < 0.01 and a significant within group comparison for animals maintained on the HFD; ***p < 0.001. E While on time-restricted feeding, there is a main effect in testing time of day with more distance traveled occurring during the inactive phase, **p < 0.01, and a significant within group comparison for animals maintained on the HFD, ***p < 0.001. Statistics were performed using SPSS Version 29.0.1.0 All values are reported as means ± SEM and were analyzed by Multivariate ANOVA

In animals on a high-fat diet, the timing of feeding has little impact on anxious-like behavior

To determine how circadian disruption, as derived from Day-fed conditions, affects anxious-like behavior, we compared the EPM behavior of the high-fat diet-fed rats exclusively fed during the active phase (Night-fed) to the behavior of those fed exclusively during the inactive phase (Day-fed). In general, we find that the circadian timing of the high-fat diet had little impact on anxious-like behavior.

Night-fed and Day-fed rats had a similar number of closed-arm, open-arm, and center entries (Feeding Length*Diet*Feeding Time of Day; 12-hour, HFD: Day-fed vs. Night-fed, F(1,70) = 1.248, F(1,70) = 1.090, F(1,70) = 1.908 respectively, all p > 0.05, Supplemental Fig. 4). Similarly, the timing of food availability had no effect on the amount of time the rat spent in the closed- or open-arms (Feeding Length*Diet*Feeding Time of Day; 12-hour, HFD: Day-fed vs. Night-fed, all p > 0.05, Supplemental Fig. 4). However, Day-fed rats did spend more time in the center than Night-fed rats (Feeding length*Diet*Feeding time; 12-hour: HFD: Day-fed vs. Night-fed, F(1,70) = 5.701, p = 0.023; Fig. 6A). Night-fed rats, overall, traveled more than Day-fed rats (Feeding length*Diet*Feeding time; 12-hour, HFD: Day-fed vs. Night-fed, F(1,70) = 9.114, p = 0.003; Fig. 6B).

Fig. 6.

Night-fed rats have significantly lower anxious-like behaviors than Day-fed rats during both the active and inactive phase. A There is a main effect of Day-Feeding to increase center time spent in HFD-fed animals; *p < 0.05. B There is a main effect of Day-Feeding to decrease total distance traveled, **p < 0.01, and significant day-night variation within each feeding condition; *p < 0.05, **p < 0.01. C Closed-arm entries are higher during the inactive phase in both HFD Day-fed and HFD Night-fed conditions; **p < 0.01. D Center entries are higher during the inactive phase in both HFD Day-fed and HFD Night-fed conditions; *p < 0.05, **p < 0.01. Statistics were performed using SPSS Version 29.0.1.0 All values are reported as means ± SEM and were analyzed by Multivariate ANOVA

The timing of food did not override the diurnal expression of behaviors. Regardless of whether rats were Night-fed or Day-fed, they continued to exhibit greater distance traveled (Feeding length*Diet*Feeding time*Testing time of Day; 12-hour, Day-fed, HFD: AM-tested vs. PM-tested, F(1,70) = 2.345, p = 0.023, 12-hour, Night-fed, HFD: AM-tested vs. PM-tested, F(1,70) = 8.046, p = 0.047; Fig. 6B) closed-arm entries (Feeding length*Feeding time*Diet*Testing time of day; 12-hour, Day-fed, HFD: AM-tested vs. PM-tested, F(1,70) = 5.531, p = 0.02, 12- hour, night-fed, HFD: AM-tested vs. PM-tested, F(1,70) = 16.696, p = 0.02 ; Fig. 6C), center entries (Feeding length*Feeding time*Diet*Testing time of day; 12-hour, Day-fed, HFD: AM-tested vs. PM-tested, F(1,70) = 6.064, p = 0.014, 12-hour, Night-Fed HFD: AM-tested vs. PM-tested, F(1,70) = 10.129, p = 0.002, Fig. 6D) during their inactive phase, consistent with their behavior during ad libitum feeding.

Exposure impacts behavioral habituation to the EPM

Generally, animals exhibit a decrease in anxious-like behaviors with repeated exposure to the EPM5. In this study, rats were exposed to the EPM four times: twice while fed ad libitum and twice during time-restricted feeding. While testing was counterbalanced between active and inactive phases, all time-restricted feeding occurred after ad libitum conditions (Fig. 1). In general, we observe a decrease in anxious-like behaviors with repeated exposure to the EPM, consistent with habituation, in both high-fat and chow-fed rats.

Rats spent more time in the open-arms (Main effect of Time: p = 0.003, Fig. 7A) and less time in closed-arms (Main effect of Time p = 0.0012, Fig. 7B) with repeated exposure to the EPM. Additionally, diet interacts with increased trial exposure, such that animals spend less time in the closed-arms with increased trial testing (Time x Diet, p = 0.0129, Fig. 7B). Time spent in the center increases with repeated exposure (Main effect of time p = 0.0065; Fig. 7C). Closed-arm entries and center entries did not exhibit any main effect of time or diet or interaction effects (Fig. 7D and E). Diurnal variation is maintained throughout the experiment (Main effect of Phase p < 0.0001; Fig. 7F). Together these data indicate that anxious-like behaviors decrease with repeated EPM trials with little to no influence from diet.

Fig. 7.

Repeated exposure to the EPM impacts habituation. A Overall, open-arm time illustrates a main effect of time with no significant main effect of diet or interaction effect. B Regardless of diet, time spent in the closed-arms has a main effect of time, p = 0.0012. Additionally, there is an interaction effect, Time x Diet, p = 0.0129. Closed-arm time decreases in chow animals from the 1st to the 3rd trial: *p = 0.0195; from the 2nd to the 3rd trial: *p = 0.0236; and from the 3rd to 4th trial: *p = 0.0327. In HFD-fed animals, closed-arm time decreases from the 1st to the 3rd trial: #p = 0.0465; and from the 1st to the 4th trial: #p = 0.0104. C Center time spent shows only a main effect of time, p = 0.0065. D There are no significant interaction or main effects in closed-arm entries when doing trial by trial comparisons. E There are no significant interaction or main effects in center entries. F There is a main effect of phase in closed-arm entries when phase is accounted for within the trial comparisons, p < 0.0001 but no interaction effects of Time x Diet. Statistics were performed in Prism Version 10. All values are reported as means ± SEM and were analyzed by Multivariate ANOVA

Discussion

Our data indicate that the time of testing significantly affects the presentation of anxious-like behavior in the EPM but not the NBT. The absence of active–inactive phase differences in the NBT may be a limitation of the extended testing window used in the present study, as prolonged exposure can promote habituation and potentially mask chronic anxiety-like behaviors. Given the lack of significant effects in the NBT, subsequent experiments focused on the EPM. Within the EPM, more anxious-like behavior occurs during the rats’ inactive phase, consistent with the clinical literature [1,3]. Moreover, these day-night differences are largely impervious to high-fat feeding, and while time-restricted feeding may reduce overall anxious-like behavior it cannot reverse diurnal differences in anxious-like behavior.

As supported by prior literature [53], we find that rats exhibit diurnal variation in anxious-like behaviors in the EPM. Rats exhibit a relative decrease in closed-arm and center entries during their active phase. This is in contrast with other studies that illustrate a peak of anxious-like behavior occurring during the inactive phase [53] or no observable peak at all [54]. This discrepancy may be due to variations in test design. In particular, a peak in anxious-like behavior occurring during inactive phase was concluded by experimenters testing rats at varying ages starting at 1 month all the way through 18 months of age on the EPM. Whereas the study detecting no peak in anxious-like behaviors was conducted in juvenile rats using a social exploration assay as opposed to EPM. Indeed, age, the type of anxiety assessment [55,56], and other factors such as environmental lighting [57–61], can significantly impact daily expression of anxious-like behavior. While locomotor activity can influence arm entries, it cannot fully explain the observed behavioral patterns. If increased locomotion were the sole contributor, proportional changes would be expected across all arm measures, yet open-arm entries did not consistently parallel total movement. Furthermore, during time-restricted feeding we find reduced closed-arm entries without altered total distance traveled, supporting the conclusion that our findings reflect changes in anxiety-related behavior rather than general activity levels. Lastly, the extended 15-minute EPM duration used in this study may reduce sensitivity to open-arm exploration, which is typically most pronounced during the initial minutes of exposure. Consequently, shorter duration assays may capture additional novelty or exploratory effects that were not detectable within our design [44]. Future studies using shorter EPM sessions may provide insights into early novelty and exploratory behaviors.

Similar to previous work in high-fat diet rodent models [62,63], our data illustrate that diet al.one does not affect anxious-like behaviors. We demonstrate that while there is a strong diurnal variation in anxious-like behaviors, the recorded behaviors of rats on a high-fat diet do not differ from those on a chow diet. It is far from clear how obesity affects anxious-like behaviors. There is contradicting research on whether high-fat diets reduce anxious-like behavior. Some researchers have demonstrated that a high-fat diet increases general anxious-like behaviors [64] while others observed a relative decrease in anxious-like behaviors when rats where on a high-fat diet [65,66]. Although our data indicate that diet al.one did not affect anxious-like behavior, prior studies have reported anxiolytic effects of high-fat diet feeding [65]. One potential explanation for these differing findings is variation in animal strain, as behavioral responses to high-fat diets may differ between Long-Evans and Wistar rats. In addition, animals in the present study were maintained on a high-fat diet for an extended duration relative to other studies which employed relatively shorter exposure. Prolonged high-fat diet exposure may attenuate or eliminate transient anxiolytic effects observed following shorter-term dietary interventions. Overall, these differences may be due to varying models, distinct diets, and testing time discrepancies. For example, genetically obese-prone rats fed a high-fat/high-carbohydrate diet exhibit increased anxious-like behaviors across both active and inactive phases [64], suggesting that excessive obesity may exacerbate anxiety presentation. In contrast, research reporting a relative decrease in anxious-like behaviors utilized a high-fat/high-sugar diet. Furthermore, duration of diet exposure may also affect anxiety presentation, where 8 weeks of exposure appear insufficient to affect anxious-like behaviors, but prolonged exposure past 15 weeks increases anxious-like behaviors [22]. Together this highlights that differences in diet composition, duration of time on a high-fat diet, and genetic susceptibility may yield varying results in anxious-like behaviors.

In agreement with a previous study [67], we show that corticosterone levels decrease as body weight increases. While we do not find any direct correlation between body weight and anxious-like behaviors in the present study, others have found that caloric restriction alongside intermittent fasting lowers anxious-like behavior in the EPM concomitant with weight loss [67].

Time-restricted feeding can mitigate the metabolic consequences of circadian disruption [47,68]. Therefore, we sought to determine the extent to which time-restricted feeding might impact rhythms in anxious-like behavior. We found that limiting high-fat diet feeding to a 12-hour feeding window decreased overall anxious-like behavior without affecting day-night variation. Interestingly, exclusively feeding the high-fat diet during the inactive period did not override the natural day-night differences in anxious-like behavior. This may implicate non-metabolic, yet clock-related circuitries involved in anxious-like behaviors. Interestingly, in contrast to high-fat fed rats, chow-fed rats lost their diurnal variation in anxious-like behaviors when exclusively Night-fed. Since chow-fed rats consume most of their calories at night under ad libitum conditions, this anxiolytic effect may be due to a combination of habituation and/or indirect limitation of caloric intake. However, because body weight was not measured following initiation of time-restricted feeding, we are unable to specifically determine how a potential loss in body weight relates to anxiety-like behavior.

This experiment was principally designed to examine diurnal differences in anxiety-like behaviors and assess restricted feeding and timing of feeding on high-fat diet-fed animals. Time-restricted feeding was used to explore how aligning food to the active phase influences rhythms in anxiety expression. However, the present design limits our ability to disentangle the effects of time-restricted feeding from habituation. We acknowledge the possibility that the observed decrease in anxious-like behaviors seen in time-restricted feeding, may reflect habituation rather than the effects of time-restricted feeding per se. For example, we observe that regardless of diet composition, 12-hour restricted feeding increased the amount of time spent in the EPM center. Given that these assessments occurred after ad libitum testing (which included two EPM exposures), this increase in center time likely reflects partial habituation to the EPM. Additional studies are required to clarify the role of time-restricted feeding on anxious-like behaviors.

Our data has some unique characteristics that should be considered. The present study used aged rats exposed to prolonged high-fat diet exposure. This may lend the present study to report higher than typical anxious-like behaviors, as young rats tend exhibit less anxious-like behaviors in the EPM [69,70]. Given the advanced age of the animals, body weight may be influenced by age-related factors independent of dietary condition. Although the chow-fed rats weighed significantly less than the high-fat diet-fed rats, their body weights were nonetheless higher than those typically reported for younger adult chow-fed rats, consistent with age-related weight gain [71]. In addition, high-fat diet exposure is associated with a range of physiological alteration beyond anxiety-related behavior. Previous studies demonstrate that high-fat diet feeding induces metabolic impairments including increased glucose intolerance [72,73] which may contribute to the broader phenotype observed in these animals. Due to our interest in examining the interplay between obesity and anxiety, we elected to test only male rats because females are largely resistant to weight gain on a high-fat diet [74]. We urge future studies to specifically test for sex differences in the daily expression of anxious-like behaviors. Finally, we note that behavioral testing is itself a perturbation. Testing during the animals’ inactive (rest) phase disrupts sleep, which can increase anxious-like behavior independent of circadian biology. As a result, observed effects may reflect testing-induced circadian/sleep disruption rather than the experimental manipulation. Given the nocturnal nature of rodents, conducting behavioral assays at a single timepoint during the animals inactive (light phase) may yield misleading or incomplete results [75]. As research increasingly acknowledges the role of circadian rhythms in shaping behavior, it is critical to account for time-of-day effects when assessing anxiety. Incorporating circadian variation into experimental design will provide a more comprehensive understanding of the temporal dynamics that influence anxiety.

Importantly, these findings extend beyond animal research and have clinical implications. Anxiety symptoms in humans also fluctuate across the day. Assessing anxiety at only one time of day in a clinical setting may overlook periods of heightened anxiety. Specifically, assessing anxiety in clinically obese patients during both their active and inactive period and better aligning food to the active, day-time period, may help to mitigate anxious behaviors. Recognizing time-of-day variability can help clinicians better characterize symptom severity, tailor intervention timing, and improve diagnostic assessments. By demonstrating the importance of circadian timing in anxious-like behavioral testing in rodents, our results underscore the broader need to integrate time-of-day into preclinical and clinical research focused on understanding anxiety behaviors.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

A.R. designed and coordinated the execution of the experiment, conducted the experiments, analyzed and interpreted the data, and wrote the manuscript. E.T. performed data collection, analysis, and interpretation, and prepared Figure 3. B.M.G. assisted with experimental design, data collection and manuscript editing. D.M.A. supervised the project, assisted in data analysis, data interpretation, and manuscript revisions. All authors reviewed and approved the final version of the manuscript.

Funding

Marquette University.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.