2.7 4.

Article

Spatial Learning of Individual

Cichlid Fish and Its Effect on Group
Decision Making

Jiaxing Long and Shijian Fu

Topic Collection
Behavioral Ecology of Aquatic Animals
Edited by
Prof. Dr. Fang Wang and Prof. Dr. Shijian Fu

[Link]
 animals
Article
Spatial Learning of Individual Cichlid Fish and Its Effect on
Group Decision Making
Jiaxing Long and Shijian Fu *

Chongqing Key Laboratory of Animal Biology, Chongqing Normal University, Chongqing 401331, China;
2020110513015@[Link]
* Correspondence: shijianfu9@[Link]

Simple Summary: Learning and memory abilities and their roles in group decision-making have
important ecological relevance in routine activities such as foraging and anti-predator behaviors in
fish species. The aims of this study were to explore the spatial learning ability of juvenile cichlids
(Chindongo demasoni), and the influence of heterogeneity of memory information among group
members on group performance in a six-arm radiation maze. The spatial performance of individual
fish improved and reached a stable level on the fifth day of training, and the memory of the space
task is kept after 11 days of detraining. The spatial performance of heterogeneous groups composed
of members with different memory information were found to change linearly with the increase
of the proportion of trained members. This indicates that cichlids can obtain associative learning
information through training, and it seems unlikely that cichlids’ group behavior is determined by
minority members in foraging context.

Abstract: Learning and memory abilities and their roles in group decision-making have important
ecological relevance in routine activities such as foraging and anti-predator behaviors in fish species.
The aims of the present study were to explore individual spatial learning abilities of juvenile cichlids
(Chindongo demasoni) in a foraging context, and to explore the influence of heterogeneity of memory
Citation: Long, J.; Fu, S. Spatial
information among group members on group performance in a six-arm radiation maze. In the context
Learning of Individual Cichlid Fish of an association between landmarks and food, learning ability was evaluated by the speed and
and Its Effect on Group Decision accuracy of reaching the arm with food during seven days of reinforcement, and memory retention
Making. Animals 2022, 12, 1318. was tested at intervals of 2, 5, 8 and 11 days of detraining. Then, the speed and accuracy of an
[Link] eight-member group with different proportions of memory-trained fish were measured. Both speed
and accuracy of individual fish improved significantly and linearly in the first five days of training
Academic Editor: Tyrone
Lucon-Xiccato
and leveled off between five and seven days, with values 60% shorter (in speed) and 50% higher (in
accuracy) compared to those of the first day. Neither speed nor accuracy showed any decrease after
Received: 11 March 2022 11 days of detraining, suggesting memory retention of the spatial task. When measured in a group,
Accepted: 19 May 2022
the speed and accuracy of the majority of the group (more than half) in reaching the arm with food
Published: 21 May 2022
changed linearly with an increasing ratio of trained members. This shows that cichlids can acquire
Publisher’s Note: MDPI stays neutral associative learning information through a training process, and group behavior of cichlids seems
with regard to jurisdictional claims in not likely be determined by a minority of group members under a foraging context.
published maps and institutional affil-
iations. Keywords: cichlids; memory retention; learning; six-arm maze; group decision

Copyright: © 2022 by the authors.

1. Introduction
Licensee MDPI, Basel, Switzerland.
This article is an open access article In recent years, research in cognitive science has been largely aimed at the in-depth
distributed under the terms and study of the brain [1,2], but the study of animal learning and memory has also become a
conditions of the Creative Commons focus of attention as an important part of cognitive science [3,4]. Learning and memory
Attribution (CC BY) license (https:// are among the most basic functions of the animal brain. The former is the process by
[Link]/licenses/by/ which animals acquire new information and knowledge, whereas the latter is the process
4.0/). of saving and reading the acquired information [5]. Associative learning is the learning

Animals 2022, 12, 1318. [Link] [Link]

Animals 2022, 12, 1318 2 of 10

of an association between two stimuli or events [6,7], whereas operant conditioning is a

type of associative learning usually occurring through repeated reinforcement of a certain
behavior [8]. Operant conditioning is often investigated in animal individuals placed in
diverse devices such as shuttle boxes, T-mazes, or plus mazes using either simple response
discrimination (right vs. left) or a visual cue (e.g., reward) marking the correct choice [9].
For example, researchers found that trained fifteen-spined sticklebacks (Spinachia spinachia)
and corkwing wrasse (Crenilabrus melops) can find food more efficiently in radiation mazes
by learning through the link between food with visual spatial cues [10].
Foraging behavior is one of the basic behaviors that animals rely on for survival, and
the spatial learning and memory of animals (especially those living in complex habitats) can
help improve their foraging efficiency, and ultimately, improve the individual’s survival
fitness [11,12]. Animals can solve a spatial task by learning an association between a
response and a reward (response strategy) or they can learn an association between a given
place and a reward (place strategy) [13]. For example, studies in zebrafish (Danio rerio)
in plus mazes found that trained zebrafish can locate the food arm through association
learning [14]. Recently, using T-mazes as arenas, a comparison study in four fish species
found a profound difference in spatial learning and memory abilities at both inter- and
intra-species levels [15]. More recent studies in cichlids such as p. taeniatus found that
the spatial learning and memory ability tested in the maze might be altered by rearing
condition and gender [16–18]. In this experiment, cichlids that usually live in complex
habitats were selected as the experimental model, and associative learning between food
and a landmark was studied in the arena of a six-arm maze. The first aim of the present
study was to test whether operant conditioning could be formed under such conditions by
evaluating the speed and accuracy of reaching the landmark arm with food located behind
the landmark during a week of training. The second aim of the present study was to test
the memory retention of the possible operant conditioning by evaluating the speed and
accuracy after different periods (2 to 11 days) of detraining. To the best of our knowledge,
this is the first study of operant conditioning through a landmark and foraging behavior to
test the memory retention in cichlids in a six-arm radiation maze.
The phenomenon where animals gather in groups and cooperate is called group
behavior, which is the survival strategy gradually evolved by animals in the process of
natural selection [19–22]. It has been suggested that communication within the group
members speeds up the transmission of information, hence improving cognition ability
and increasing the efficacy of foraging [23,24]. An animal group usually needs to make
collective decisions when faced with different options while foraging, often deciding be-
tween multiple spatially distributed options [25]. Collective decision-making research has
received widespread attention in recent years [26–28]. Based on these studies, researchers
have made certain theoretical breakthroughs in understanding group decision-making
mechanisms through mathematical modeling [13]. For example, researchers have proposed
two group decision-making mechanisms: meritocratic leadership decisions and consensus
decisions. The main feature of the former is that members of the group can disproportion-
ately affect the group decision-making through leadership [7,29]. However, on the contrary,
the exchange of information and members’ contributions to the group behavior in a con-
sensus decision appear to be more even and equal [20]. Although important theoretical
breakthroughs have been made in group decision-making, the results of investigations in
fish species are quite controversial [30,31], and the underlying mechanism needs further
investigation. Recently, some researchers proposed a radiative six-arm maze device system,
which provides a possible paradigm of discrete alternatives in a series, to be faced by the
individuals in an animal group and is an ideal device for group decision-making theory
verification and related experiments [7]. Thus, the third aim of this study was to inves-
tigate the decision-making mechanism of cichlids by evaluating the speed and accuracy
of a fish group comprising different ratios of trained and untrained (i.e., informed and
non-informed) members. Because it is difficult to identify individuals within the group,
the speed and accuracy of the fish group were evaluated based on the majority (i.e., more
Animals 2022, 12, 1318 3 of 10

than half the members of the group). We anticipated that both variables might improve in
groups with a presence of trained fish as a consequence of communication among group
members. Furthermore, we assumed that if the speed and accuracy improved profoundly
in a group with the presence of only a few trained members but showed less improvement
with the increase in trained members, the collective decision making might be affected by a
minority of group members. On the contrary, if group performance was linearly correlated
with the ratio of informed individuals, the group behavior would be more determined
equally within the group members.
Cichlidae belong to the class Osteichthyes, of order Perciform. The family is an ideal
candidate for fish behavior research [23] as these fish are beautiful in appearance, easy to
raise, and have complex social group behaviors [24]. In this experiment, juvenile Demasoni
(Chindongo demasoni) from Lake Malawi were selected as the experimental animals. The
pilot experiment showed that this species possesses a good learning and memory capacity.

2. Materials and Methods

2.1. Acclimation and Handling of Experimental Fish
The cichlids used in the experiment were purchased in October 2021 from the Mashi
Aquarium in Shapingba District, Chongqing. The purchased experimental fish were
sterilized with a 1–2% concentration of sodium chloride solution and placed in a self-
purification circulating and temperature-controlled rearing system (the volume of each
rearing unit was about 20 L; the length was 35 cm, the width was 20 cm and the height
was 28 cm) for 14 days, and the water temperature was maintained at 25 ± 0.5 ◦ C during
acclimation. The water for acclimation was tap water that was aerated for 24 h. During the
acclimation period, air was continuously added to the water with an air pump to ensure that
the dissolved oxygen in the water body was maintained above 7 mg L−1 . Experimental fish
were fed with commercial feed once daily at 9:00, and after one hour, the remaining food
and fish excrement and feces were siphoned out; the daily water exchange was about 10%
of the total water body. After acclimation, 360 healthy juveniles with a mass of 2.52 ± 0.59 g
and a body length of 4.60 ± 0.70 cm were selected as the experimental subjects.

2.2. Experimental Setup and Behavioral Measurement Protocol

The experimental device was a modified six-arm maze (Figure 1), designed according
to earlier studies [25]. Each arm of the device was a 20 cm × 40 cm cuboid structure. The
side-length of the regular hexagon in the center was also 20 cm, and the water depth of the
maze was 5 cm. The experimental light sources were six fluorescent lamps that ensured
the uniform illumination of the shooting site. The plastic plant was used as landmark, i.e.,
the associative information for food in the present study [32]. During the experiment, an
arm was randomly selected, and a plastic plant placed within it. Approximately 5 mg of
commercial feed was added to the end of the same arm. In this way, the experimental
fish could not directly see the food, and could only reach and consume it after they swam
through the plastic plant. The location of food along with the plastic plant was changed
each day to avoid the effect of spatial orientation ability on analysis of the results [33–35].
Before the start of each experiment, the experimental fish were transferred from the rearing
system using an opaque cylindrical adaptor and placed in a blank arm spaced one arm
away from the arm with food for 10 min adaptation according to a pilot experiment. We
placed a cylinder at the mouth of the arm they were in to prevent the experimental fish
from seeing the arm with food in advance. Then, the adaptor was removed, and a camera
(LogiCapture 920) placed 1 m directly above the equipment was used for continuous
shooting at 15 frames per second for 10 min.
Animals 2022, 12, 1318 4 of 10

‐
Figure 1. Schematic drawing of the radial six-arm maze used in the present study.

2.3. Experimental Scheme

2.3.1. Experiment 1: Training and Memory Test
Training Process
Sixty individuals were randomly selected. Each day at approximately 9:00, 12:00
and 15:00, fish were transferred to the maze individually three times. The activity of
experimental fish was recorded for 10 min after a 10-min pre-acclimation period : (see
detail in aforementioned measurement protocol). The training process lasted for seven‐
consecutive days. The arm in which the artificial‐ plants and
‐ food were placed was changed
daily to avoid the influence of spatial memory on the behavior of the experimental fish [33].
Among the 60 trained fish, 5 individuals showed no improvement in performance (see
parameters for detail) and so their training process was terminated between the second
and fourth days.

Memory Retention Test

After the seven days of training, 48 individuals were randomly selected and the
activity was recorded exactly the same as the training process, once again, after 2, 5, 8 and
11 days of detraining (12 repetitions at each time point). After the first training process, the
‐
fish were reared individually, meaning each fish could be identified during the training
process and memory retention test.

2.3.2. Experiment 2: Group Decision Test ‐

After the memory retention test, heterogeneous trained groups were formed by differ-
ent ratios of trained and non-trained fish (0:8, 1:7, 3:5, 5:3, 7:1 and 8:0) That is, 0:8 refers
to 0 nontrained fish and 8 trained fish, 1: 7 refers to 1 nontrained fish and 7 trained fish,
and so on. We conducted eight repetitions for each ratio. Trained fish were used repeatedly‐
after they underwent two rounds ‐ of measurement (0:8 to 8:0), whereas nontrained fish
were used only once. In brief, eight individuals as a group comprised a specific number of
trained or nontrained fish, which were randomly selected. Then, all group members were‐
immediately moved into the second left or right arm next to the arm with food to make
sure that the fish could not see the landmark before they swam out of the arm. ‐
The water temperature and other conditions during the experimental tests were the
same as those in the acclimation period, and all experimental fish were tested after fasting
for two days.

2.4. Experimental Parameters

The parameters of the speed and accuracy in reaching the arm with food were used
to measure the fish’s cognitive performance. The speed to reach the arm with food was
calculated as the period from the beginning of the observation to when the fish went
through the plants (observed by the experimenter via video). Accuracy was calculated as
the proportion of experimental fish to reach the arm with food the first time (i.e., crossing
the landmark) among all fish test members.
Animals 2022, 12, 1318 5 of 10

Because experimental fish cannot be identified individually, and they seldom form a
complete group (i.e., all group members in only one arm), for the group decision-making
measurement, a majority group was defined as more than half of the group members
(i.e., five fish in the present study) located in any arm of the maze. The speed and
accuracy of the majority of the group were calculated exactly the same as we did for
individual measurements.

2.5. Data, Statistics and Analysis

The data were first inputted to Excel 2003, and then the SPSS 26.0 software was used
for statistical analysis. The training effect was tested using a linear mixed model using the
fish ID as the random factor and the training days as the main factor. Memory retention was
tested by a linear mixed model using the fish ID as the random factor and the detraining
period and treatment (seven days trained vs. detrained) as the main factors. The effect of
group composition on the speed of the majority in the arm with food was tested by one-way
analysis of variance (ANOVA). Then, Duncan’s multiple comparisons were used to analyze
the statistical differences between the groups. The relationship between group composition
and accuracy for the majority of the group was analyzed by linear regression. The data are
expressed as the “mean ± standard error”, and the significance level is p < 0.05.

3. Results
3.1. The Training Effect and Memory Retention of Experimental Fish
3.1.1. Training Effect
Training had a significant effect on both the speed (F1,57 = 208.01, p < 0.001) and
accuracy rates (F1,57 = 816.53, p < 0.001) based on the results of the linear mixed model.
The speed, i.e., time taken to first arrive at the arm with food, decreased gradually and
significantly from 22.53 s to 9.17 s during the first five days of training (Figure 2a) (p < 0.05).
However, there was no further decrease from the fifth to seventh days of training. Similarly,
the accuracy rate, i.e., proportion of experimental fish to first arrive in the arm with food
rather than arm without food, increased from 44.25% (which is much higher than 13.3%, the
theoretically calculated value of one-sixth) to 60.34% during the first five days of training
(there was no significant difference in the first four days, although the values gradually
increased) (Figure 2b) (p < 0.05), whereas there was no further increase from the fifth to
seventh days of training.
a) b)
80 a
30 a ab
Accuracy rate (%)

25 bc abc
60 c bc
b c
Speed (s)

20 bc bc
15 cd 40
d d
10 20
5
0 0
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Training period (d) Training period (d)

Figure 2. The speed (a) i.e., time when fish first arrived at arm with food, and accuracy rate (b)
i.e., proportion of fish that first entered the arm with food rather than an arm without food, for the
foraging task in the maze during the training period (means ± S.E., n = 55). Note: a, b, c, d letters
suggest a significant difference of the variables after different days of training (p < 0.05).

3.1.2. Memory Retention

Neither speed nor accuracy rate of fish after the seventh day of training and those
with 2 to 11 days of detraining showed any significant difference (Table 1).
‐
Animals 2022, 12, 1318 6 of 10

Table 1. Results of the memory retention test, i.e., comparison of speed and accuracy rate after 7 d
of training with those for the same fish following a different period of detraining based on a linear
mixed-model analysis (parameter data are presented as means S.E.).

Statistical Analysis
Detraining Period 2d 5d 8d 11 d Treatment Period Interaction
Effect Effect Effect
F1,43 = 0.297 F3,43 = 0.912 F3,43 = 0.506
Training 7.33 ± 0.94 8.86 ± 2.54 8.17 ± 0.84 8.06 ± 0.92
Speed (s) p = 0.589 p = 0.175 p = 0.680
F3,42 =0.164 F3,42 = 0.473 F3,42 = 0.473
Detraining 10.82 ± 6.18 6.42 ± 1.56 6.42 ± 1.38 5.73 ± 1.39
p = 0.920 p = 0.703 p = 0.703
F1,86 = 1.691 F3,86 = 0.349 F3,86 = 1.237
Accuracy Training 50 ± 0.15 69.44 ± 0.14 55.56 ± 0.15 69.44 ± 0.13
p = 0.197 p = 0.790 p = 0.301
rate (%)
F3,42 = 0.564 F3,42 = 0.780 F3,42 = 0.113
Detraining 61.11 ± 0.14 69.44 ± 0.14 55.56 ± 0.15 66.67 ± 0.14
p = 0.642 p = 0.512 p = 0.593

3.2. The Effect of Group Composition on the Decision Making of Experimental Fish
The speed with which the majority of the fish arrived at the arm with food de-
creased gradually and significantly with an increasing proportion of trained fish (Figure 3a)
(p < 0.05). The among-group variation also decreased profoundly, as can be seen from the
S.E. of each value. The accuracy rate of the majority, for first entering the arm with food
rather than the arm without food, increased in line with the proportion of trained fish
(y = 0.808x + 20.032, R2 = 0.970, p < 0.001), from 12.5% in the 0:8 group to 100% in the 8:0
group (Figure 3b).

250
a ab
200

150
Speed (s)

abc
100
bc
c
50 c
0
0 12.5 37.5 62.5 87.5 100
Proportion of trained fish (%)

(a)

100
y = 0.808x + 20.032
90
80 R² = 0.970, P < 0.001
70
60
Accuracy (%)

50
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
Proportion of trained fish (%)

(b)
Figure 3. The speed (n = 8) (a) and accuracy (b) of the majority of the fish in the group to reach the
arm with food, with different proportions of trained individuals. Note: a, b, c, d letters suggest a
significant difference of the variables after different days of training (p < 0.05).
Animals 2022, 12, 1318 7 of 10

4. Discussion
4.1. Learning and Memory Abilities of Cichlids
The experimental fish showed learning ability through an association between a
landmark and food; the speed and accuracy of cichlids entering the arm with food improved
profoundly with an increase in training time. The speed and accuracy improved steadily
over the first five days of training, suggesting that such operant conditioning could be
reinforced by increased training repetition. Similar results have been documented in fish
species such as zebrafish [36]. However, no further increase in either speed or accuracy was
noted from five to seven days, suggesting there is a threshold on this cognitive task. The
speed and accuracy rose with the training process and finally reached a relatively stable
threshold [37]. Eventually, the time required to find the arm with food shortened by 59.30%
(i.e., from 22.53 s to 9.17 s), while the accuracy increased by 50.65% (i.e., from 44.25% to
66.67%). Notably, the improvement in speed was far more profound than that of accuracy.
A possible reason for a more profound change in speed might be that the experimental fish
behaved more bravely as they became familiar with the arena and training process. It is
worth pointing out that the 40% accuracy on the first day is far higher than our theoretically
calculated value (i.e., one-sixth, 13.3%). This suggests that cichlids prefer to enter the
arm with plastic plants even without food information, possibly as plants act as shelter in
natural habitats [38]. Thus, the accuracy rate had less room for improvement during the
training process. Although this preference does have an influence, it has no effect on our
experimental results. With the increase in training days, the association is still reinforced as
the speed of individual fish becomes faster and faster and the accuracy becomes higher
and higher.
The results of the memory retention test in the present study showed that cichlids
can hold on to cognitive ability throughout training for at least 11 days, as no difference
was found in either speed or accuracy after 2 to 11 days of detraining. Similar results have
been documented in other animals, including fish species. Researchers have proposed
that birds and mammals can retain their learning and memories for several months to
years [39–41]. There are also extensive examples showing that the learning and memory
capabilities of fish are impressive [42]. It has been found that gobies (Bathygobius soporator)
can retain their memory for 40 days [43], while tilapia (Sarotherodon galilaeus) retain their
memory of sound stimuli for up to six months [44]. Some researchers have suggested
that cleaner fish (Labroides dimidiatus) can remember events such as being captured by nets
for at least 11 months [45]. During their lives, fish will face many situations. They need
to choose the most effective way to make use of their living environment [46], learn to
explore and identify high-quality food resources [47], resolve conflicts with competitors,
find suitable mates, and avoid being tracked by enemies [48]. All these require them to
make appropriate responses to the situation at such times. Learning and memory play
an important role in these behaviors of fish; therefore, the abilities of both learning and
memory are consistent with fish’s adaptation to the environment and meeting their survival
and reproduction needs.

4.2. Effects of Group Composition on Foraging Performance and the Possible

Decision-Making Mechanism
The results showed that the speed and accuracy of the majority of the fish group
entering the arm with food increased in line with an growing proportion of trained group
members. This suggests that the behavior of the cichlid group might be decided equally by
the group members, i.e., cichlids might adopt democracy in group decision making under a
foraging context. Cichlids are a social species, living in high-predation habitats. A previous
study found that the decision making of fish populations from high-predation habitats
is more likely to be equally determined [49]. However, contrary to our expectations, the
present study found that cichlids’ group foraging takes longer than that of individuals
during the training process. The study suggests that behavioral interaction among group
members might not always lead to an increased foraging efficiency in cichlids. Thus, the
Animals 2022, 12, 1318 8 of 10

linear increase in performance with an increase in the number of trained fish, estimated by
the majority of a group, in either speed or accuracy might be possible due to the fact that
trained fish will be in the majority. Identifying each fish member to acquire the trajectory
of each trained and nontrained group member in future can help better understand how
the composition affects the group decision [50]. Nevertheless, the linear relationship
between group composition and performance in the present study suggests that the group
performance of cichlids might not be determined by a minority of group member with a
specific threshold in a foraging context. It is worth noting that, by using the same six-arm
maze, our observation found that time for the majority of the cichlid group to enter an
arm with shelter after a simulated predator stimulation was much shorter than those of
individual cichlids. This suggests that an ecological context other than foraging might be
more appropriate to study the collective-decision mechanism for cichlids.
In conclusion, the present study found that cichlids show an impressive associative
learning ability when performing a foraging task in a six-arm maze. Such cognitive
ability might be ecologically important for cichlids living in complex habitats with high
heterogeneity of food abundance. The foraging behavior of cichlids in the group seems
unlikely be determined by a minority of group members in a foraging context. Further
studies in other ecological contexts might yield interesting results.

Author Contributions: J.L., S.F.; Conceptualization: S.F.; Methodology: S.F., J.L., Formal Analysis:
J.L., S.F.; Resources: J.L.; Data curation: J.L.; Writing— Original Draft: J.L., S.F.; Writing—Review and
Editing: S.F.; Funding acquisition: S.F., J.L. All authors have read and agreed to the published version
of the manuscript.
Funding: This study was funded by the National Natural Science Foundation of China (31670418).
Institutional Review Board Statement: This study complied with the current law of the country
in which it was performed and was approved by the Animal Care and Use Committee of the Key
Laboratory of Animal Biology of Chongqing (Permit Number Zhao-20,210,611-01). The study was
performed in strict accordance with the recommendations in the Guide for the Care and Use of
Animals at the Key Laboratory of Animal Biology of Chongqing, China. All applicable international,
national and/or institutional guidelines for the care and use of animals were followed.
Informed Consent Statement: Not applicable.
Data Availability Statement: Data are available upon request from the corresponding author.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Kosse, C.; Burdakov, D. Natural hypothalamic circuit dynamics underlying object memorization. Nat. Commun. 2019, 10, 1–8.
[CrossRef] [PubMed]
2. von Bastian, C.C.; Belleville, S.; Udale, R.C.; Reinhartz, A.; Essounni, M.; Strobach, T. Mechanisms underlying training-induced
cognitive change. Nat. Rev. Psychol. 2022, 1, 30–41. [CrossRef]
3. Zu, G.; Sang, H.; Zhou, H.; Zhang, F.; Wu, X.; Wang, A.; Zhang, M. Flexible attention system: Appearance time of split attention
changes in accordance with the task difficulty level. Atten. Percept. Psychophys 2022. [CrossRef]
4. Lu, E.; Hu, X. Image super-resolution via channel attention and spatial attention. Appl. Intell. 2021, 52, 2260–2268. [CrossRef]
5. Johnston, M.V.; Alemi, L.; Harum, K.H. Learning, memory, and transcription factors. Pediatric Res. 2003, 53, 369–374. [CrossRef]
[PubMed]
6. Pearce, J.M. Animal Learning and Cognition: An Introduction; Psychology Press: Hove, UK, 2013.
7. Vega-Trejo, R.; Boussard, A.; Wallander, L.; Estival, E.; Buechel, S.D.; Kotrschal, A.; Kolm, N. Artificial selection for schooling
behaviour and its effects on associative learning abilities. J. Exp. Biol. 2020, 223, jeb235093. [CrossRef]
8. Anthouard, M. A Study of Social Transmission in Juvenile Dicentrarchus Labrax (Pisces, Serranidae), in an Operant Conditioning
Situation. Behaviour 1987, 103, 266–275. [CrossRef]
9. Blaser, R.E.; Vira, D.G. Experiments on learning in zebrafish (Danio rerio): A promising model of neurocognitive function. Neurosci.
Biobehav. Rev. 2014, 42, 224–231. [CrossRef]
10. Hughes, R.N.; Blight, C.M. Two intertidal fish species use visual association learning to track the status of food patches in a radial
maze. Animal Behaviour 2000, 59, 613–621. [CrossRef]
11. Odling-Smee, L.C.; Boughman, J.W.; Braithwaite, V.A. Sympatric species of threespine stickleback differ in their performance in a
spatial learning task. Behav. Ecol. Sociobiol. 2008, 62, 1935–1945. [CrossRef]
Animals 2022, 12, 1318 9 of 10

12. Braithwaite, V.; McGown, A. Caregivers’ emotional well-being and their capacity to learn about stroke. J. Adv. Nurs. 1993,
18, 195–202. [CrossRef] [PubMed]
13. Gibson, B.M.; Shettleworth, S.J. Place versus response learning revisited: Tests of blocking on the radial maze. Behav. Neurosci.
2005, 119, 567. [CrossRef] [PubMed]
14. Margarette, S.; Robert, G. Associative learning in zebrafish (Danio rerio) in the plus maze. Behav. Brain Res. 2010, 207, 99–104.
15. Zhu, Y.R. Studies on Spatial Learning and Memory of Some Freshwater Fishes; Institute of Hydrobiology, Chinese Academy of Sciences:
Wuhan, China, 2007.
16. Hesse, S.; Sandmann, S.; Bakker, T.C.; Thünken, T. Impact of social rearing-environment on performance in a complex maze in
females of a cichlid fish. Behav. Process. 2019, 167, 103915. [CrossRef] [PubMed]
17. Salena, M.G.; Singh, A.; Weller, O.; Fang, X.X.; Balshine, S. Rapid spatial learning in cooperative and non-cooperative cichlids.
Behav. Processes 2021, 194, 104550. [CrossRef] [PubMed]
18. Wallace, K.J.; Hofmann, H.A. Equal performance but distinct behaviors: Sex differences in a novel object recognition task and
spatial maze in a highly social cichlid fish. Anim. Cogn. 2021, 24, 1057–1073. [CrossRef] [PubMed]
19. Seeley, T.D.; Camazine, S.; Sneyd, J. Collective decision-making in honey bees: How colonies choose among nectar sources. Behav.
Ecol. Sociobiol. 1991, 28, 277–290. [CrossRef]
20. Shaw, E. The schooling of fishes. Sci. Am. 1962, 206, 128–141. [CrossRef]
21. Delcourt, J.; Poncin, P. Shoals and schools: Back to the heuristic definitions and quantitative references. Rev. Fish Biol. Fish. 2012,
22, 595–619. [CrossRef]
22. Domeier, M.L.; Colin, P.L. Tropical reef fish spawning aggregations: Defined and reviewed. Bull. Mar. Sci. 1997, 60, 698–726.
23. Couzin, I.D. Collective cognition in animal groups. Trends Cogn. Sci. 2009, 13, 36–43. [CrossRef] [PubMed]
24. Lindstedt, E.R.; Hamilton, I.M. Variation in social information use: The influences of information reliability and mass on decision
making in a group-living fish Gambusia affinis. J. Fish Biol. 2013, 82, 2095–2103. [CrossRef] [PubMed]
25. Delcourt, J.; Miller, N.Y.; Couzin, I.D.; Garnier, S. Methods for the effective study of collective behavior in a radial arm maze.
Behav. Res. Methods 2018, 50, 1673–1685. [CrossRef]
26. Wade AS, I.; Ramnarine, I.W.; Ioannou, C.C. The effect of group size on the speed of decision making depends on compromise
and predation risk across populations in the guppy Poecilia reticulata. Behaviour 2020, 157, 1173–1192. [CrossRef]
27. Mehlis, M.; Thünken, T.; Bakker, T.; Frommen, J.G. Quantification acuity in spontaneous shoaling decisions of three-spined
sticklebacks. Anim. Cogn. 2015, 18, 1125–1131. [CrossRef] [PubMed]
28. Miller, N.; Garnier, S.; Hartnett, A.T.; Couzin, I.D. Both information and social cohesion determine collective decisions in animal
groups. Psychol. Cogn. Sci. 2013, 110, 5263–5268. [CrossRef]
29. Warburton, K. Learning of Foraging Skills by Fishes; John Wiley Sons, Ltd.: Hoboken, NJ, USA, 2007.
30. Sumpter DJ, T.; Krause, J.; James, R.; Couzin, I.D.; Ward, A.J.W. Consensus decision making by fish. Curr. Biol. 2008, 18, 1773–1777.
[CrossRef] [PubMed]
31. Bisazza, A.; Butterworth, B.; Piffer, L.; Bahrami, B.; Petrazzini, M.E.M.; Agrillo, C. Collective enhancement of numerical acuity by
meritocratic leadership in fish. Sci. Rep. 2014, 4, 1–5. [CrossRef]
32. Zhu, Y.R.; Liu, H.Z. Influence of landmark in spatial learning and memory in Rhodeus ocellatus. J. Huazhong Agric. Univ. 2015,
34, 5.
33. Murphy, K.E.; Pitcher, T.J. Predator attack motivation influences the inspection behaviour of European minnows. J. Fish Biol. 1997,
50, 407–417. [CrossRef]
34. Williams, F.E.; White, D.; Messer, W.S. A simple spatial alternation task for assessing memory function in zebrafish. Behav.
Processes 2002, 58, 125–132. [CrossRef]
35. Croy, M.I.; Hughes, R.N. The influence of hunger on feeding behaviour and on the acquisition of learned foraging skills by the
fifteen-spined stickleback, Spinachia spinachia L. Anim. Behav. 1991, 41, 161–170. [CrossRef]
36. Miletto Petrazzini, M.E.; Pecunioso, A.; Dadda, M.; Agrillo, C. The Impact of brain lateralization and anxiety-like behaviour in an
extensive operant conditioning task in zebrafish (Danio rerio). Symmetry 2019, 11, 1395. [CrossRef]
37. Healy, S.D.; Braham, S.R.; Braithwaite, V.A. Spatial working memory in rats: No differences between the sexes. Proc. R. Soc.
London. Ser. B Biol. Sci. 1999, 266, 2303–2308. [CrossRef]
38. Nonato FA, S.; Michelan, T.S.; Freitas, P.V.; Maia, C.; Montag, L.F.D.A. Heterogeneity of macrophyte banks affects the structure of
fish communities in flooded habitats of the Amazon Basin. Aquat. Ecol. 2021, 55, 215–226. [CrossRef]
39. Boeckle, M.; Bugnyar, T. Long-term memory for affiliates in ravens. Curr. Biol. 2012, 22, 801–806. [CrossRef]
40. Sosna, M.M.G.; Twomey, C.R.; Bak-Coleman, J.; Poel, W.; Daniels, B.C.; Romanczuk, P.; Couzin, I.D. Individual and collective
encoding of risk in animal groups. Proc. Natl. Acad. Sci. USA 2019, 116, 20556–20561. [CrossRef]
41. Bruck, J.N. Decades-long social memory in bottlenose dolphins. Proc. R. Soc. B Biol. Sci. 2013, 280, 20131726. [CrossRef]
42. Brown, C. Fish intelligence, sentience and ethics. Anim. Cogn. 2015, 18, 1–17. [CrossRef]
43. Aronson, L.R. Further studies on orientation and jumping behavior in the gobiid fish, Bathygobius soporator. Ann. N. Y. Acad. Sci.
1971, 188, 378–392. [CrossRef]
44. Zion, B.; Barki, A.; Grinshpon, J.; Rosenfeld, L.; Barki, A. Retention of acoustic conditioning in St Peter’s fish Sarotherodon galilaeus.
J. Fish Biol. 2011, 78, 838–847. [CrossRef] [PubMed]
Animals 2022, 12, 1318 10 of 10

45. Triki, Z.; Bshary, R. Long-term memory retention in a wild fish species Labroides dimidiatus eleven months after an aversive event.
Ethology 2020, 126, 372–376. [CrossRef]
46. Odling-Smee, L.; Braithwaite, V.A. The influence of habitat stability on landmark use during spatial learning in the three-spine
stickleback. Anim. Behav. 2003, 65, 701–707. [CrossRef]
47. Warburton, K. Learning of foraging skills by fish. Fish Fish. 2003, 4, 203–215. [CrossRef]
48. Kelley, J.L.; Magurran, A.E. Effects of relaxed predation pressure on visual predator recognition in the guppy. Behav. Ecol. Sociobiol.
2003, 54, 225–232. [CrossRef]
49. Herbert-Read, J.E.; Rosén, E.; Szorkovszky, A.; Ioannou, C.C.; Rogell, B.; Perna, A.; Ramnarine, I.W.; Kotrschal, A.; Kolm, N.;
Krause, J.; et al. How predation shapes the social interaction rules of shoaling fish. Proc. R. Soc. B Biol. Sci. 2017, 284, 20171126.
[CrossRef]
50. Stienessen, S.C.; Parrish, J.K. The effect of disparate information on individual fish movements and emergent group behavior.
Behav. Ecol. 2013, 24, 1150–1160. [CrossRef]