Behavioral Lateralization and Boldness Traits Across Eight Teleost Fish Species

Abstract

Understanding inter-species differences in behavioral lateralization and exploration patterns is crucial for advancing the study of animal behavior. In this study, we standardized the experimental procedure to minimize methodological variability and examined the behavioral responses of eight fish species (Girardinus falcatus, Poecilia reticulata, Paracheidon axelrodi, Kriptoterus bichirris, Hyphessobrycon megalopterus, Danio rerio, Corydoras aeneus, and Xenopoecilius sarasinorum) in a novel, circular environment. We focused on boldness-related measures (latency to explore and freezing time) and motor activity (circular vs. linear swimming). Significant inter-species differences were observed in the boldness measures. Fish also showed a preference for circular swimming over linear swimming. However, no lateralization bias (clockwise vs. counterclockwise) was detected in any species. These findings may establish a baseline for future research on the spontaneous behaviors of eight teleost species and offer valuable insights for the design of future behavioral studies focusing on these species.

1. Introduction

Behavioral lateralization refers to the asymmetric expression of cognitive functions. The most prominent example in our species is handedness, where approximately 90% of the population is right-handed [1]. Behavioral lateralization, however, is not a human prerogative and has been reported in other mammals, birds, amphibians, reptiles, and fish (reviewed in [2,3]). This asymmetrical distribution of cognitive functions is thought to provide advantages in terms of cognitive efficiency and energetic costs as it avoids the duplication of neural circuits for the same cognitive abilities [4], reduces inter-hemispheric conflict in the control of cognitive functions [5], and permits greater efficiency in dual tasks [6].

However, while in humans, there is evidence of a robust population bias, at least in hand preferences, assessing whether there is consistency at a population level within a vertebrate group is always challenging. Different studies use varying approaches to assess the degree and direction of behavioral lateralization. Even when adopting the same type of test (for instance, a detour test requiring turning left or right in a T-maze), the size/shape of the apparatus and the experimental protocol differ among studies, thus making cross-species comparisons difficult [7,8].

Teleost fish are known to display various types of lateralized behaviors, for instance when turning left or right, when monitoring a predator or prey with one eye over the other, or when analyzing a social companion (reviewed in [9]). One of the most efficient methodological approaches to test behavioral lateralization in basal vertebrates, including fish, consists of observing spontaneous behavior in a novel circular environment. An asymmetrical proportion of time swimming in a clockwise (or counter-clockwise) direction is assumed to reflect the existence of a bias in swimming in one direction [10,11,12]. However, although lateralization biases frequently occur at individual levels by fish [5,13], the existence of lateralization at the population level often remains unclear. It is also uncertain whether the same degree and direction of brain lateralization in a rotational preference test occurs in all teleost fish.

Additionally, in this type of rotational test, subjects are commonly inserted into a novel, unfamiliar environment, a condition that may often lead to freezing as an anti-predator strategy. To collect data on behavioral lateralization, cognitive ethologists often delay the observation time by a few minutes, thus allowing subjects to acclimatize to the novel context. The time preceding the beginning of the motor activity is still relevant for a broader comprehension of the animal models under investigation in behavioral lateralization studies, but the time in which the animal remains motionless is often not considered for the analyses or is confined to studies on boldness. Similarly, recording the amount of freezing during explorative behavior is another variable that might help to describe inter-species differences in free swimming. Both the latency to explore the environment and the amount of freezing are commonly used as a measure of boldness [14,15,16].

In the present work, we selected some of the most commonly used species in laboratory studies and observed them in the rotational preference tests. In particular, we tested zebrafish (Danio rerio), guppies (Poecilia reticulata), goldbelly topminnows (Girardinus falcatus), glass catfish (Kryptopterus bicirrhis), Corydoras albinus (Corydoras aeneus), cardinal tetra (Paracheirodon axelrodi), black phantom tetra (Hyphessobrycon megalopterus), and Sarasin’s minnow (Xenopoecilus sarasinorum). Most of these species are freshwater fish originating from tropical or subtropical regions, particularly in South America and Southeast Asia, and inhabit slow-moving or standing waters such as streams, ponds, and swamps [17,18,19]. D. rerio, P. reticulata, and G. falcatus are surface or midwater dwellers often found in shallow, vegetated waters, where they feed on small invertebrates and algae. In contrast, bottom-dwelling species like C. aeneus exhibit benthic foraging behavior, adapted to scavenging in sandy or muddy substrates. K. bicirrhis is unique in this group due to its semi-transparent body and preference for dimly lit, slow-flowing waters. Tetras like P. axelrodi and H. megalopterus are highly social, shoaling species from the Amazon basin, adapted to acidic, tannin-rich waters with low light. X. sarasinorum, a lesser-known species endemic to Sulawesi, Indonesia, inhabits lake environments and exhibits livebearing reproduction, a trait shared with P. reticulata and G. falcatus. Despite sharing general traits such as small body size and omnivorous diets, these species demonstrate ecological diversity in terms of habitat preference, social behavior, and trophic strategies, making them valuable comparative models in laboratory research.

To reduce the methodological variability and permit a fine cross-species comparison, we used the same apparatus and procedure for the eight species. Fish were individually inserted into a circular environment. Latency to explore was recorded, as well as the linear/circular swim. Spontaneous clockwise and counter-clockwise directions of swimming were also observed. Additionally, we analyzed the freezing time of each species in an attempt to assess the proportion of swimming in the whole observation time. This comprehensive cross-species comparison aims to elucidate inter-species difference in boldness traits and the distinctive behavioral lateralization patterns of commonly used teleost models, thereby enhancing our understanding of species-specific traits and facilitating the more informed selection and interpretation of fish models in future lateralization/personality studies.

2. Materials and Methods

2.1. Subjects

A total of 165 freshwater fish were tested, averaging at 21 for each species (see

Table 1. Number of subjects tested for each species.

Species	Sample Size
Girardinus falcatus	21
Poecilia reticulata	20
Paracheidon axelrodi	20
Kryptoterus bichirris	23
Hyphessobrycon megalopterus	19
Danio rerio	20
Corydoras aeneus	21
Xenopoecilius sarasinorum	21

). Subjects were adult sexually mature individuals maintained in groups of 18–25 (contained in 150-liter tanks, 60 × 40 × 40 cm) at the Laboratory of Comparative Psychology of the University of Padova. An 18 W fluorescent light was provided above each stock tank (photoperiod was 14:10 h light/dark), and each tank also has air filters, natural gravel, and live plants. The water temperature was set to 25 ± 1 °C. Fish were fed twice daily with commercial food flakes in the morning and with live brine shrimp (Artemia salina) in the afternoon.

The study complies with all laws of the country in which it was performed (Italy) and was approved by the local ethics committee of the University of Padova (12/2021). The authors complied with the ARRIVE guidelines.

2.2. Apparatus and Procedure

The apparatus consisted in a white cylinder (33 cm in diameter and 29 cm in height) made with plastic material. The floor was also made with the same white material. The tank was filled with 15 cm of water (temperature: 25 ± 1 °C). Additionally, 18 W fluorescent lamps were placed above the tank to light the experimental area. A video camera was suspended about 1 m above the test tank and used to record the position of the subjects.

The procedure was similar to that adopted in previous studies investigating rotational swimming [10,11]. Subjects were singly inserted into the experimental tank by a fish net. The observation period started after 15 min of acclimatation. During the acclimatation period, we recorded the latency to explore the environment. The observation period lasted 15 min; in this phase, we divided fish swimming into three different activities: (a) circular clockwise swimming, (b) circular counterclockwise swimming, and (c) linear swimming. Freezing was also recorded during this period. At the end of the observation period, fish were captured by the net and inserted in a stock tank similar to the one used previously.

All tests were video-recorded. Video were used to identify our dependent variables in the following way:

(1)

Latency to explore the environment was defined by the time necessary to change the position of their head (in any direction) for at least 50% of the subjects’ body size during the acclimation period. Such a movement must occur within 4 s. This excluded the possibility that minimal, slow movements might be considered the beginning of the explorative behavior.

(2)

Swimming activity:

(2a)

Circular clockwise swim: total time spent swimming in the clockwise direction (usually in correspondence of the circular walls) during the observation period.

(2b)

Circular counterclockwise swim: total time spent swimming in the counterclockwise direction (usually in correspondence of the circular walls) during the observation period.

(2c)

Linear swim was defined by the total time spent swimming in a non-circular direction (e.g., from the north–south direction, east–west direction, etc.; Figure 1).

(3)

Freezing behavior was defined by the total time in which fish remained motionless during the observation period. The criteria for defining motionless behavior were identical to the ones used for the latency to explore the environment.

Figure 1. Examples of clockwise swimming (left), counterclockwise swimming (center), and linear swimming (right).

Figure 1. Examples of clockwise swimming (left), counterclockwise swimming (center), and linear swimming (right).

2.3. Statistical Analyses

Inter-species differences across three behavioral variables were examined as follows:

Latency to Explore the Environment: Latency was analyzed using a one-way analysis of variance (ANOVA), with species as the between-subjects factor.

Swimming Activity: Swimming activity was assessed using a mixed-design ANOVA, with motion type (circular clockwise, circular counterclockwise, and linear swimming) as the within-subjects factor and species (eight levels) as the between-subjects factor. Population biases in the circular swimming direction was assessed with one-sample t-tests conducted on the proportion of clockwise swimming for each species. The chance level was set to 0.5. We also calculated Bayes factors (one sample) to assess the strength of evidence in favor of the null hypothesis (no side bias) versus the alternative hypothesis.

Freezing Behavior: Freezing behavior during the observation period was analyzed using a one-way ANOVA, with species as the between-subjects factor. We also correlated the latency to explore the environment with freezing during the explorative behavior (Pearson correlation).

In the event of significant main effects, post hoc comparisons were performed using the Least Significant Difference (LSD) test. All statistical analyses were conducted using SPSS version 23.0.

3. Results

3.1. Latency to Explore the Environment

A one-way ANOVA revealed a significant main effect of species on latency to initiate movement (F(7, 164) = 23.494, p < 0.001, partial eta squared η² = 0.512; Figure 2). Post hoc analyses indicated that P. axeldori and H. megalopterus required significantly more time to begin exploring the novel environment compared to other species (see

Table 2. Post hoc analyses of species with respect to latency.

Multiple Comparisons (LSD)	Dependent Variable	Latency to Explore
Species    vs.	Species	Sig.
Girardinus	Paracheirodon	<0.001
	Kryptopterus	0.858
	Hyphessobrycon	<0.001
	Danio	0.804
	Corydoras	0.214
	Poecilia	0.883
	Xenopoecilus	0.971
Paracheirodon	Kryptopterus	<0.001
	Hyphessobrycon	0.209
	Danio	<0.001
	Corydoras	<0.001
	Poecilia	<0.001
	Xenopoecilus	<0.001
Kryptopterus	Hyphessobrycon	<0.001
	Danio	0.939
	Corydoras	0.148
	Poecilia	0.978
	Xenopoecilus	0.887
Hyphessobrycon	Danio	<0.001
	Corydoras	<0.001
	Poecilia	<0.001
	Xenopoecilus	<0.001
Danio	Corydoras	0.141
	Poecilia	0.920
	Xenopoecilus	0.831
Corydoras	Poecilia	0.170
	Xenopoecilus	0.201
Poecilia	Xenopoecilus	0.911

for detailed comparisons).

3.2. Swimming Activity

The mixed-design ANOVA revealed a significant main effect of species (F(7, 157) = 19.283, p < 0.001, partial η² = 0.462), a main effect of swim (Greenhouse–Geisser-corrected, F(1.794, 281.628) = 321.528, p < 0.001, partial η² = 0.672), and a significant species × swim interaction (Greenhouse–Geisser-corrected, F(12.557, 281.628) = 6.083, p < 0.001, partial η² = 0.213, Figure 3). Post hoc comparisons (

Table 3. Post hoc analyses of species with respect to the swimming activity.

Multiple Comparisons (LSD)	Dependent Variable	Swimming Activity
Species     vs.	Species	Sig.
Girardinus	Paracheirodon	0.026
	Kryptopterus	<0.001
	Hyphessobrycon	<0.001
	Danio	0.105
	Corydoras	0.107
	Poecilia	<0.001
	Xenopoecilus	0.009
Paracheirodon	Kryptopterus	0.248
	Hyphessobrycon	<0.001
	Danio	<0.001
	Corydoras	0.517
	Poecilia	0.001
	Xenopoecilus	0.727
Kryptopterus	Hyphessobrycon	<0.001
	Danio	<0.001
	Corydoras	0.067
	Poecilia	0.025
	Xenopoecilus	0.418
Hyphessobrycon	Danio	<0.001
	Corydoras	<0.001
	Poecilia	0.005
	Xenopoecilus	<0.001
Danio	Corydoras	0.002
	Poecilia	<0.001
	Xenopoecilus	<0.001
Corydoras	Poecilia	<0.001
	Xenopoecilus	0.313
Poecilia	Xenopoecilus	0.003

) revealed substantial inter-species variability in swimming patterns, with distinct preferences for circular versus linear motion across species. We then focused on circular swimming and did not find any significant population bias in the proportion of clockwise direction for any species (one-sample t-tests, see

Table 4. Proportion of clockwise swim as a measure of side bias. Inferential and Bayesian analyses support the idea of a lack of population bias for any species.

Species	Prop. of Clockwise Swim (Mean ± Std. Dev)	t-Tests, p-Value and Cohen’s d	One Sample Bayes Factor
Girardinus falcatus	0.548 ± 0.190	t(20) = 1.172, p = 0.255, d= 0.256	3.153
Poecilia reticulata	0.437 ± 0.141	t(19) = −2.007, p = 0.059, d = −0.449	1.006
Paracheidon axelrodi	0.535 ± 0.191	t(17) = 0.778, p = 0.447, d= 0.183	4.193
Kryptoterus bichirris	0.504 ± 0.057	t(22) = 0.340, p = 0.737 d= 0.071	5.913
Hyphessobrycon megalopterus	0.559 ± 0.235	t(11) = 0.862, p = 0.407, d= 0.249	3.304
Danio rerio	0.543 ± 0.129	t(19) = 1.488, p = 0.153, d= 0.333	2.132
Corydoras aeneus	0.478 ± 0.076	t(20) = −1.302, p = 0.208, d= −0.284	2.729
Xenopoecilius sarasinorum	0.437 ± 0.044	t(20) = −1.414, p = 0.173, d= −0.309	2.386

).

3.3. Freezing Behavior

A one-way ANOVA showed a significant main effect of species on freezing behavior (F(7, 164) = 19.295, p < 0.001, partial η² = 0.462, Figure 3). The post hoc analysis (

Table 5. Post hoc analyses of species with respect to the freezing activity.

Multiple Comparisons (LSD)	Dependent Variable	Freezing
Species     vs.	Species	Sig.
Girardinus	Paracheirodon	0.030
	Kryptopterus	<0.001
	Hyphessobrycon	<0.001
	Danio	0.112
	Corydoras	0.102
	Poecilia	<0.001
	Xenopoecilus	0.009
Paracheirodon	Kryptopterus	0.218
	Hyphessobrycon	<0.001
	Danio	<0.001
	Corydoras	0.568
	Poecilia	<0.001
	Xenopoecilus	0.666
Kryptopterus	Hyphessobrycon	<0.001
	Danio	<0.001
	Corydoras	0.067
	Poecilia	0.025
	Xenopoecilus	0.422
Hyphessobrycon	Danio	<0.001
	Corydoras	<0.001
	Poecilia	0.005
	Xenopoecilus	<0.001
Danio	Corydoras	0.002
	Poecilia	<0.001
	Xenopoecilus	<0.001
Corydoras	Poecilia	<0.001
	Xenopoecilus	0.310
Poecilia	Xenopoecilus	0.003

) indicated that H. megalopterus and P. reticulata spent significantly more time freezing than other species.

3.4. Correlation Between Latency and Freezing

When data were pooled across species, a significant positive correlation was found between latency to explore and time spent freezing (Pearson r = 0.398, p < 0.001; Figure 4), suggesting that individuals who delayed initiating movement were also more likely to exhibit prolonged freezing behavior during exploration.

4. Discussion

The present study aimed to examine inter-species differences in boldness and behavioral lateralization among fish during the exploration of a novel environment. To minimize the methodological variability commonly observed in the literature, we standardized both the apparatus and procedures across all species. Our findings indicate significant variation in motor activity, with a significant preference for circular swimming over linear swimming. This preference aligns with the cylindrical shape of the environment and corroborates the well-established literature on the thigmotactic response in fish, wherein they tend to avoid the center and prefer moving along the boundaries of an arena when exploring unfamiliar spaces [20,21]. However, when assessing the lateralization of the swimming direction (clockwise vs. counterclockwise), no significant bias was observed in any species, indicating no evidence of lateralization at either the population or species level [22]. This result is partially corroborated by the alternative statistical approach adopted to assess behavioral lateralization: Bayes factors showed moderate evidence in favor of the null hypothesis (no population bias) in four species, namely G. falcatus, P. axelrodi, K. bichirris, and H. megalopterus. Anecdotal evidence for the null hypothesis comes from Bayes factors of the other species, leaving open the possibility (especially for P. reticulata) that lateralization bias might exist in some species. It is worth noting that our subjects were observed only once, thus limiting the possibility to assess individual-level lateralization. We call for future studies to assess also whether inter-species differences might exist in the proportion of individuals with a robust lateralization bias.

In contrast, substantial inter-individual variation was observed in boldness measures. Freezing behavior is typically associated with defensive immobility in response to perceived threats, especially in shyer individuals [23]. The species exhibited significant differences in freezing behavior, suggesting a population-level bias along the boldness–shyness continuum, with some species more readily engaging with novel, brightly lit environments than others. H. megalopterus, in particular, exhibited the greatest proportion of freezing behavior. This species, characterized by a dark body coloration, has previously been shown to exhibit a pronounced scototactic response, preferring darker environments more than other teleosts, both individually and in groups [19]. Future studies are necessary to directly assess whether light-level tolerance is correlated to the amount of freezing in these species. In our experimental apparatus, the absence of shelter and the high luminosity of the environment might have elicited more pronounced freezing behavior in H. megalopterus than in other species.

Species also differed in their latency to move, with H. megalopterus again demonstrating the longest latency. White et al. [24] showed that multiple measures of behavior are necessary to adequately quantify boldness in fish, because these behavioral measures were not interchangeable. Here we attempted to see whether the two measures of boldness correlate one to another. To facilitate the correlational analysis, we pooled data from all species, although we acknowledge the inter-species variability. A positive correlation emerged between latency to explore and the amount of freezing observed throughout the observation period. Although this result is intuitive, we believe it may hold some methodological implications for future studies. Specifically, as initial latency to explore is predictive of subsequent freezing behavior, screening subjects for their response to the environment at the onset of an experiment might help exclude individuals with longer latencies, potentially reducing the inclusion of uncooperative subjects.

Our study can offer novel contributions to the literature on behavioral lateralization and boldness in teleost fish. First, unlike many previous investigations that focused on single species or employed heterogeneous experimental designs, this study adopts a standardized methodological approach across eight phylogenetically diverse fish species, thereby reducing procedural variability and enhancing cross-species comparability. Second, it provides comparative data on boldness-related behaviors—specifically, latency to explore and freezing time—across eight species, which is rare in the existing literature. Third, the research introduces a novel analytical focus on motor patterns within a circular environment, revealing a general preference for circular over linear swimming, an aspect that has been overlooked in previous studies. Finally, although further investigation is necessary before drawing a firm conclusion, the lack of lateralization bias in all tested species may add a critical dimension to our understanding of spontaneous locomotor asymmetries, challenging assumptions about the ubiquity of lateralized behavior in fish and calling for more nuanced investigations into the ecological or evolutionary factors driving such traits.

5. Conclusions

This study, which was exploratory in nature, aimed to assess inter-species differences in the exploratory behavior of fish in a novel, circular environment, with no a priori predictions. Contrary to previous findings [25,26], we found no evidence of lateralization biases in the direction of swimming (clockwise vs. counterclockwise) nor any inter-species differences in this respect, suggesting that teleost fish may not significantly differ in the direction/degree of lateralization bias at a population level. In contrast, significant inter-species variation was observed in measures of boldness, including latency to move and the amount of freezing behavior. These measures were positively correlated, such that longer latencies to explore were predictive of greater freezing behavior later in the experiment. We hope that our study offers a useful starting point for future investigations into the spontaneous behaviors of common teleost species, particularly regarding boldness, and may help inform the design of subsequent experiments involving these species.