Maturation of shoaling behavior is accompanied by changes in the dopaminergic and serotoninergic systems in zebrafish

Abstract

The zebrafish has been one of the preferred vertebrate model organisms of developmental biology, and is becoming an important research tool for behavioral neuroscience and behavior genetics. A prominent feature of zebrafish is their strong shoaling tendency. Most recently, the first paper investigating the development of shoaling in zebrafish demonstrated that a few days after hatching zebrafish do not shoal, but that shoaling tendency gradually increases during development. The current paper investigates this phenomenon using the nearest neighbor distance, a measure most frequently employed for the quantification of shoal cohesion in fish. We demonstrate that shoal cohesion increases with age, while thigmotaxis, “wall hugging”, does not show a consistent age-dependent change. The mechanisms underlying the maturation of shoaling are unknown. HPLC analysis of whole brain extracts finds the concentration of dopamine, DOPAC, serotonin, and 5-HIAA normalized to total brain protein weight to increase with age. Although the behavioral and neurochemical results are only correlative at this point, they may open a new avenue into the investigation of the mechanisms and development of social behavior in zebrafish.

INTRODUCTION

The zebrafish has been proposed as a useful model organism to investigate a range of neurological and neuropsychiatric conditions, e.g. addiction (Gerlai, Fernandes, & Pereira, 2009; Mathur & Guo, 2009), autism, stress, and anxiety (Mathur & Guo, 2009). As such, zebrafish are increasingly used in behavioral studies (Gerlai, 2010), and are argued to be a suitable model for high-throughput mutation (forward genetic) screening (Gerlai, 2010; Kokel et al., 2010). Most previous screening efforts focused on embryonic characteristics because the zebrafish embryo is transparent and develops rapidly (Brittijn et al., 2009). In addition to genetics, zebrafish have also started to be employed in chemical screens (Tsang, 2009). Briefly, the zebrafish is argued to be an excellent research tool with which a variety of human diseases may be modeled and the mechanisms of complex functions, including brain functions, may be studied.

Numerous aspects of embryonic development, including brain development, of zebrafish have been well characterized (Keller, Schmidt, Wittbrodt & Stelzer, 2008; Ross, Parrett & Easter, 1992). Furthermore, an increasing number of studies have also investigated adult neural function and the behavior of adult zebrafish (Guo, 2004). Interestingly, however, there are almost no studies that investigated developmental changes that may occur from after early embryogenesis to adulthood. A notable exception is that of Engeszer et al (2007) who investigated the development of shoaling and the potential plasticity of shoal mate preference in zebrafish. These authors showed that shoaling starts early (appears first at post-flexion stage) and shoal preference depends upon visual access to shoal mates, leading to a preference that manifests at later stages (juvenile) towards the color variant the developing fish were exposed to.

Notably, a 5 dpf (day post fertilization) old zebrafish fry has reached free swimming stage and is about 3–4 mm long and is almost completely transparent (Rubinstein, 2003). A sexually mature young adult (about 90 dpf old) is 3–4 cm long, fully colored, and exhibits a complex behavioral repertoire. We noticed that 5–10 dpf old zebrafish appear scattered and do not form tight groups, shoals, whereas one of the most robust behaviors that is consistently observed in adult zebrafish is shoaling. In a recent paper we have investigated this phenomenon and demonstrated that shoaling indeed develops with age (Buske & Gerlai, 2011). Importantly, in this latter study we monitored and quantified shoaling behavior in freely moving zebrafish groups and not as a response of a single isolated fish to the sight of conspecifics behind a glass barrier, as was conducted by Engeszer et al (2007). We quantified shoaling by measuring inter-individual distances among all shoal members. This measure calculates all distances among all shoal members but is not frequently utilized in the analysis of shoaling in fish because it is sensitive to the number of shoal members. A measure that is independent of shoal size is the nearest neighbor distance which takes into account only how far the nearest neighbor of a fish is and calculates this distance for each shoal member. Arguably, this measure is biologically relevant as fish may not perceive or respond to all members within a shoal and may only focus on the nearest neighbors when making motor responses (Pitcher & Parrish, 1993). Thus, in the current study we utilize this measure and quantify the changes in the nearest neighbor distance as zebrafish develop from age 10 days post fertilization (dpf) to 75 dpf.

What mechanisms may underlie the age-dependent changes in shoaling are not known. There may be numerous neuroanatomical, neurophysiological, gene expression and biochemical/molecular changes that accompany the development of shoaling and which of these may represent causal relationships may be difficult to know in advance. To start the investigation of such mechanisms we decided to analyze potential developmental changes in neurochemical levels of the zebrafish brain.

Previously, we have shown that shoaling, i.e. access to conspecifics is rewarding for zebrafish (Al-Imari & Gerlai, 2008). The dopaminergic system has been shown to subserve several functions in the brain one of which is reward (e.g. Cannon & Bseikri, 2004). Thus as a start we decided to measure dopamine levels as well as the level of the dopamine metabolite, DOPAC. Serotonin has also been shown to be involved in a range of behavioral functions, for example, aggression and fear (e.g. Popova, 2008). Given that fear (antipredatory behavior) is one of the driving forces that has been shown to underlie shoaling in multiple fish species and that aggression (intra-specific agonistic behavior) is also expected to significantly affect shoaling responses (Magurran, 1990), we decided to also quantify levels of serotonin, and the levels of one of its metabolites, 5HIAA.

Admittedly, the descriptive results presented here will not allow one to determine what mechanisms may explain the development of shoaling in zebrafish. However, they represent the first step, a correlative analysis, which we hope will allow us to formulate working hypotheses about such mechanisms to be investigated in the future.

METHODS

Animals and Housing

Zebrafish (Danio rerio) of the AB strain were utilized for the experiments. The fish originated from the Zebrafish International Research Centre (ZIRC) (Eugene, Oregon) and have bred in-house for 3 generations by the time of the experiments. All fish were kept in groups of ten throughout their development. That is, each shoal tested in the behavioral test comprised of fish that were from the same holding tank. The unit of statistical analysis of the behavioral data was the shoal, and the sample sizes (n ranging from 8 to 10 across the different ages) represent the number of shoals tested. The handling and maintenance of fish used for neurochemical analysis were the same as for those tested behaviorally. Gender could not visually be determined in a reliable fashion at the time of testing but we have found the gender ratio to be 50:50% on average in this strain. The methods of fish breeding, maintenance and behavioral experimentation followed those described previously (Buske & Gerlai, 2011).

Upon hatching, the fish were housed in groups of ten in 2.8l Plexiglas aquaria (standard Aquaneering Inc. (San Diego, CA) zebrafish tanks). At 15 days post fertilization the aquaria were connected to a recirculating filtration aquaculture rack system which had a mechanical, biological, and activated carbon filter as well as a UV sterilizing unit (Aquaneering Inc., San Diego, Ca, USA). Prior to being connected to the aquaculture rack system, water changes were performed every other day to avoid debris and toxic waste buildup in the aquaria. Water was maintained at 27°C. The system water used on the rack as well as during the development and testing of the fish was reverse osmosis purified and was supplemented with 60mg/l Instant Ocean Sea Salt to achieve water chemistry appropriate for zebrafish.

Zebrafish were kept at a 12h light/12h dark cycle with lights on at 7:00 h and off at 19:00 h. The fluorescent lights illuminated the racks from the ceiling providing an even diffuse light to all tanks. All fish were fed twice daily with Larval Artificial Plankton 100 (particle size below 100 mm, ZeiglerBros, Inc., Gardners, PA, USA) until two weeks post fertilization, after which animals were fed twice daily with nauplii of brine shrimp (Artemia salina) until they were four weeks old. After this, the developing fry were fed a mixture of flake food (Tetramin Tropical fish flake food, Tetra Co, Melle, Germany) and powered spirulina (1 part, Jehmco Inc., Lambertville, NJ, USA).

Quantification of shoaling

Fish were netted as a group and released in the center of the test tank. Fish were allowed to habituate to the test tank for 120 seconds and subsequently the recording session started. Each behavioral observation lasted for 335 seconds during which the fish were video-recorded using an overhead video camera (JVC Everio Hard Drive GZ-MG750BU). After the session, the group was returned to its home tank. All behavior recording sessions were conducted between 10:00 and 16:00 h, i.e. in the middle of the light cycle of the fish. The tank size was kept proportional to the body length of the growing fish (the linear dimensions of the tank were 28x of the average body length of the given age group tested) as recommended by Gallego & Heath (1994), Masuda et al., (2003) and Vogel (2008), and as employed for zebrafish previously (Buske & Gerlai, 2011).

Shoaling behavior of ten groups (each group containing 10 fish) was quantified repeatedly as the fish developed, a longitudinal developmental analysis. The behavioral test was conducted at 10, 12, 16, 20, 27, 30, 34, 36, 38, 40, 42, 44, 46, 48, 50, 53, 55, 57, 59, 61, 64, 67, 69, 71, 73 and 75 dpf. Using a custom software application developed in our laboratory (Miller & Gerlai, 2008), we have sampled each video-recording and measured the nearest neighbor distance (expressed in body lengths) for every member of a given group once every 5 sec (5 sec frame resolution) for the 335 sec recording session. For each frame we averaged the obtained nearest neighbor values for the given group. This gave us one value for each of the 68 frames × 26 age groups. Analysis of this data set using a nested double repeated measure ANOVA design with frame and age as factors would have given us 68 × 26 (i.e. 1768) levels, which was beyond the capability of our statistical software application. To simplify this analysis first we conducted repeated measure ANOVAs with a factor “frame” separately for each age group. This allowed us to investigate whether there were time dependent changes within the behavioral recording session. We found no consistent within session temporal changes and thus, subsequently, averaged all the nearest neighbor values obtained for all frames. Using this averaged nearest neighbor distance data set we then conducted a repeated measure ANOVA with factor “age” to investigate whether shoaling changed across the different age groups, the main question of our behavioral study. Subsequently we investigated which age group differed from which. Because post hoc multiple comparison tests are inappropriate for repeated measure designs we decided to compare every sixth age group thereby reducing the number of comparisons and thus type one error. For this analysis we conducted repeated measure ANOVAs with age as a factor with two levels (the two age groups for every pair-wise comparison) and employed a Bonferroni correction for one tailed p value to further minimize type one error (the null hypothesis was that increasing age does not decrease the average nearest neighbor distance). In addition, we also measured the distance of each shoal member from the center of the tank, a measure of thigmotaxis that has been found to reflect novelty induced fear in mammals (e.g. Treit & Fundytus, 1988) and in zebrafish as well (Champagne et al., 2010). We analyzed the effect of age using repeated measure ANOVA on this behavioral measure similarly to the way we analyzed nearest neighbor distances.

HPLC analysis

For the analysis of neurochemical levels a separate set of fish were tested. A longitudinal developmental study cannot be conducted for the analysis of potential changes of neurochemicals because the sampling procedure requires sacrificing the fish. Thus, this analysis was a cross sectional developmental study. Zebrafish of different ages were sacrificed at the same time to obtain a total of 648 brain samples. Fish were sacrificed at 15, 20, 27, 36, 38, 40, 46, 49, 54, 57, 60, 64, 66, 70, 73, 85, 90, and 96 dpf. The time of day when fish were sacrificed corresponded to that used for the behavioral analysis (i.e. between 10:00 and 16:00 h) and the time of day of sacrifice was randomly varied across the different age groups. The above age groups were chosen so as to correspond to and extend the developmental time points used in the behavioral analysis. Fish were sacrificed by decapitation, the brains were quickly dissected on ice and placed in a microcentrifuge tube which was then kept at −80°C. In order to obtain sufficient amount of tissue, 12 brains were pooled for juvenile and 3 brains were pooled for adult fish. We employed a highly sensitive HPLC method recently adopted for zebrafish (Chatterjee & Gerlai, 2009). Briefly samples were centrifuged and 5μl of the supernatant was analyzed using a BAS 460 MICROBORE-HPLC system with electrochemical detection (Bio-analytical Systems Inc., West Lafayette, IN, USA) together with a Uniget C-18 reverse phase microbore column as the stationary phase (BASi, Cat. No. 8949). Standard dopamine, DOPAC (Sigma Chemicals, St. Louis, MO, USA), serotonin and 5-HIAA (Sigma) were used to quantify and identify the peaks on the chromatographs. Dopamine, DOPAC, serotonin, and 5-HIAA levels were quantified in a single run. Data were analyzed using the SPSS statistical software package version 14. Non-repeated ANOVAs were used with AGE as the between subject factor.

RESULTS

First we have analyzed the within observation session changes in the temporal pattern of the nearest neighbor distance for every age group. This analysis showed that although the distance values fluctuate within the session, there was no statistically detectable consistent temporal change, increase or decrease of distance across the time points (video frames) sampled. That is, the distance values remained stable for the period (335 sec) of the behavioral observation session. Therefore we averaged the data obtained for the entire session and analyzed whether the average nearest neighbor distance changed as the zebrafish matured. Our results suggested that the nearest neighbor distance decreased as the fish got older, i.e. their shoal cohesion increased with age, the change appeared particularly robust between 30 and 40 dpf (figure 1). This apparent age effect was confirmed by ANOVA, which found it significant (F(25, 225) = 56.213, p < 0.001). Post hoc multiple comparison tests are not appropriate for repeated measure designs. Thus to answer the question whether older fish shoals had significantly smaller nearest neighbor distance values we compared every sixth age group in a pair-wise manner using repeated measure ANOVA (with age having only two levels) followed by a bonferroni correction with a one tailed significance test. This comparison thus reduced the number of statistical tests to 5 and allowed us to test the difference between age groups approximately separated by 11 days. ANOVA found a significant difference between 10 and 30 dpf old fish (F(1, 9) = 12.976, p < 0.05), between 30 and 42 dpf old fish (F(1, 9) = 21.484, p < 0.01), between 53 and 64 (F(1, 9) = 19.695, p < 0.01), but not between 42 and 53 (F(1, 9) = 2.361, p > 0.05), or between 64 and 75 (F(1, 9) = 4.646, p > 0.05). These results confirm that with age zebrafish form increasingly tighter shoals.

Fig. 1.

The average nearest neighbor distance (measured in body length) significantly decreases with the age of zebrafish. Mean ± S.E.M are shown. Sample sizes (n) ranged between 8 and 10. For details see Methods and Results.

The pattern of age dependent changes in the distance from the center of the tank appeared different from the above (figure 2). Although there appears to be a robust variation among age groups, there does not seem to be any consistent trend. This observation was supported by an overall repeated measure ANOVA, which found a significant age effect (F(25, 225) = 13.462, p < 0.001) but without a consistent age dependent trend (ANOVAs for the comparison of age the same group pairs as explained above, F(1, 9) < 6.375, p > 0.05).

Fig. 2.

The distance from the center of the test tank (measured in body length) does not show a consistent age-dependent change. Mean ± S.E.M are shown. Sample sizes (n) ranged between 8 and 10. For details see Methods and Results.

Analysis of the concentration of neurochemicals also revealed significant age dependent changes. Figure 3 shows the changes across age groups in dopamine levels normalized by total brain protein weight. ANOVA revealed a significant (F(15, 88) = 15.978, p<0.001) effect of age demonstrating that as fish matured, Dopamine concentration increased relative to the total weight of protein in the sample. Post hoc Tukey HSD test confirmed this and showed that fish sampled between 15 and 36 dpf exhibited the smallest Dopamine concentration values which did not significantly differ from each other (p > 0.05). 15 dpf fish showed the smallest value, which significantly differed (p < 0.05) from all fish older than 36 dpf; 20, 27, and 36 dpf fish showed significantly (p < 0.05) smaller values as compared to all fish older than 49 dpf; 38, 40, and 46 dpf fish significantly differed (p < 0.05) from fish older than 66 dpf fish, the latter age groups exhibiting the highest dopamine concentration levels.

Fig. 3.

Dopamine levels in the brain (normalized to total brain protein) significantly increase with the age of zebrafish. Mean ± S.E.M are shown. Sample sizes (n) ranged between 4 and 8. For details see Methods and Results.

Analysis of DOPAC concentrations in the zebrafish brain revealed a developmental change similar to what was found for Dopamine (Figure 4). The age-dependent increase was also found significant (ANOVA (F(15, 88) = 10.399, p<0.001). Post hoc Tukey HSD test showed no significant (p > 0.05) change between the ages 15 dpf and 46 dpf; 15, 20 and 27 dpf fish significantly (p < 0.05) differed (smaller values) from fish older than 46 dpf fish; 36 dpf fish significantly differed (p < 0.05) from fish older than 49 dpf; 38 dpf fish showed significantly smaller values compared to fish above the age of 64 dpf; and 40 dpf fish had smaller DOPAC levels (p < 0.05) compared to age groups above 57 dpf; other differences were non-significant.

Fig. 4.

DOPAC (3,4-dihydroxyphenylacetic acid) levels in the brain (normalized to total brain protein) significantly increase with the age of zebrafish. Mean ± S.E.M are shown. Sample sizes (n) ranged between 4 and 8. For details see Methods and Results.

The analysis of Serotonin levels (Figure 5) in the developing zebrafish brain showed a significant age effect (ANOVA F(15, 88) = 4.977 , p < 0.05). Tukey HSD confirmed this and revealed that up to and including age 46 dpf, fish did not significantly differ from each other (p > 0.05) but fish of this younger age range had significantly (p < 0.05) lower Serotonin levels compared to that of the older age groups.

Fig. 5.

Serotonin levels in the brain (normalized to total brain protein) significantly increase with the age of zebrafish. Mean ± S.E.M are shown. Sample sizes (n) ranged between 4 and 8. For details see Methods and Results.

Serotonin’s metabolite, 5-HIAA, exhibited a pattern of age-dependent changes similar to that of serotonin (Figure 6). ANOVA demonstrated a significant age-effect (F(15, 88) = 9.063 , p < 0.001). Post hoc Tukey HSD test showed that, in general, younger age groups had smaller 5-HIAA levels. For example, age groups 15, 20, 27, 36 and 46 were all significantly different (p < 0.05) from older age groups (except that 27, 36 and 46 dpf fish were not significantly different from 57 dpf fish, 36 dpf fish were not significantly different from 70 dpf fish, and 38 dpf fish were not significantly different from any fish).

Fig. 6.

Levels of 5-hydroxyindoleacetic acid (5-HIAA) significantly increase with the age of zebrafish. Mean ± S.E.M are shown. Sample sizes (n) ranged between 4 and 8. For details see Methods and Results.

DISCUSSION

We have previously described the ontogeny of zebrafish shoaling behavior from 7 to 120 dpf (Buske & Gerlai, 2011) using a variable called “inter-individual distance”. Inter-individual distance takes into account all distances between all members of the shoal and thus it is a precise estimate of shoal cohesion. However, because this measure is dependent upon shoal size it is less frequently used in the analysis of shoaling and other types of group forming. Dependence on shoal size is a particular problem in the analysis of shoal cohesion in natural shoals or in situations where the number of shoal members may not be precisely controlled. In shoals of larger number of individuals, the distance between shoal members of distant parts of the shoal is naturally bigger. Furthermore, some have suggested that shoaling fish may not monitor, or may not be able to monitor, their distance from all members of the shoal but rather keep their distance stable from their nearest neighbor, or neighbors, by only attending to these neighbor(s) (e.g. Pitcher & Parrish, 1993 and references therein). Thus it has been suggested that the ethologically most relevant measure of shoal cohesion may be nearest neighbor distance. In the current paper, we analyzed shoal cohesion using this measure of shoaling, the nearest neighbor distance, which is independent of shoal size. This measure also confirmed our previous observation: the distance between shoal members decreased, i.e. shoal cohesion increased, as the fish matured between their age of 10 and 75 dpf.

It is important to note that in the current study we could not dissociate age-dependent changes from other time-dependent effects in our analysis of shoal cohesion. For example, repeated exposure to the test apparatus or repeated handling may have contributed to the observed differences between fish of different ages. However, we argue this is unlikely because recently we have described similar age-dependent changes both in a longitudinal as well as in a cross sectional study, and notably in the latter fish of all age groups were tested only once. The lack of effect of repeated exposure to the test environment and the lack of effect of handling is also supported by the absence of consistent age-dependent changes in our other behavioral measure, thigmotaxis. Thigmotaxis is believed to represent a fear response in rodents (e.g. Treit & Fundytus, 1988) and also in zebrafish (Champagne et al, 2010). In case of repeated exposure induced habituation to the test environment one would expect consistent reduction of thigmotaxis, while in case of experience dependent sensitization induced by repeated handling one would expect gradual increase of thigmotaxis. However, neither was observed. It is also notable that the observed age-dependent neurochemical changes also suggest that repeated exposure plays no role given that for this analysis a non-repeated design was used.

The adaptive function of age-dependent increase of shoaling is not known, but a variety of possible hypotheses may be formulated. Based upon the known antipredatory role of shoaling it is possible that with increasing body size, zebrafish become more vulnerable to piscivores that hunt individual prey and thus shoaling represents an adaptive protective mechanism against predation (Herczeg, Gonda & Merila, 2009). Similarly, it is possible that foraging in older (larger) fish requires identification of more scattered and centralized food sources, which may be discovered more efficiently by a shoal vs. single individuals. It is also likely that close proximity to members of the other sex gains increasing significance as the fish mature and may be one of the driving forces behind increased shoaling in this cooperative spawning species. Which of the above possibilities is/are correct will require future empirical studies.

It is also important to note that the mechanisms underlying the observed developmental change in shoaling behavior are not understood. Developmental analyses, including anatomical and molecular characterization of changes usually focus on embryonic stages of zebrafish (Schweitzer & Driever, 2009; Holder & Xu, 2008), i.e. the period of development up to 5 dpf, the free-swimming stage. To the best of our knowledge, our study is the first in which some correlates of behavioral changes are analyzed beyond this developmental stage. We found significant age-dependent increases in dopamine, its metabolite DOPAC, and serotonin and its metabolite 5-HIAA. Because we expressed the amount of these neurochemicals relative to total amount of protein in the brain, the age-dependent elevation of neurochemical levels we found is unlikely to be due to the overall increase of the size of the brain in the growing fish. Therefore, we hypothesize that the increased abundance of neurochemicals we detected in older fish brains is indicative of more developed (larger number of synaptic connections) and/or more active neurotransmitter systems (larger amount of neurotransmitter synthesized and released and increased turnover of the neurotransmitter).

In general, these age-dependent neurochemical changes appear to correlate well with the observed maturation of shoaling. However, a causal link between the neurochemical and behavioral changes cannot be established at this point. The temporal resolution (number of age groups analyzed) and the power (number of samples analyzed) of the current study does not allow us to make conclusive arguments about the similarities and differences between the temporal trajectories of neurochemical and behavioral changes. For example, it appears that the levels of dopamine and DOPAC increase linearly, while serotonin and 5HIAA show a step-wise jump, a rapid increase around 40 to 50 dpf preceded and followed by a more gradual increase or steady state. The differences between the developmental trajectories of these two neurotransmitter systems are unlikely to be due to experimental error given that these neurochemicals were extracted from the same brain tissue and were analyzed in a single run at the same time using HPLC. Notably, the developmental trajectory of shoal cohesion as quantified here also shows a rapid change, an increase of shoal cohesion between 30 and 40 dpf. However, whether this particular temporal trajectory is typical of zebrafish or may have resulted from seasonal or other time-dependent environmental changes, and whether changes in the dopaminergic and serotoninergic system may explain the behavioral changes will have to be investigated in the future.

Clearly, more refined (higher resolution) temporal analyses of the age-dependent behavioral and neurochemical changes are required to provide conclusive answers to the above questions. Furthermore, whether the elevation of neurochemicals is specific to the dopaminergic and serotoninergic systems is also not known at this point. Numerous other neurochemicals may also change during development and maturation of zebrafish and these changes may also contribute to the observed age dependent shoal cohesion increase. Thus it is premature to conclude about which neurotransmitter system(s) may underlie the behavioral changes. It is also not known, although likely, that the elevated levels are specific to particular brain regions, another question that will be addressed in the future. Nevertheless, the current results demonstrating significant age-dependent behavioral and neurochemical changes now provide strong rationale for follow up studies. Also important to note that numerous molecular components other than neurotransmitter systems may need to be investigated and perhaps systematic gene expression analysis, using for example DNA microarrays, or analysis of the proteome may be required before a clear picture as to the mechanisms of the development of shoaling can emerge.

In addition to temporally (higher resolution developmental) and systematically (broader range of molecular, anatomical and neurochemical changes) more refined descriptive analyses of the observed age dependent changes, one may also suggest hypothesis driven studies that could test the mechanisms underlying these changes. Although perhaps it is too early to suggest what molecular targets one may need to investigate, it is notable that the experimenter already has a formidable tool set under his/her belt available for zebrafish. Increasing number of studies (for examples, see Sison & Gerlai, 2011) start to find that classical psychopharmacological tools, e.g. neurotransmitter receptor antagonists and agonists (small molecules) developed for mammalian (including human) species work well with zebrafish due to evolutionary conservation of the molecular targets. Thus psychopharmacological manipulation of putative targets involved in developmental changes in shoaling may be analyzed using early embryonic exposure to such small molecules. Similarly, both targeted mutagenesis as well as random mutagenesis (for a recent review see e.g. Gerlai, 2010), may be employed to manipulate potential targets involved.

Clearly, a lot has to be done before answers to the above questions may be obtained. But the first important step, the demonstration of age-dependent changes in shoaling and in the dopaminergic and serotoninergic systems in zebrafish has now been made.