Journal of Neuroscience Methods 234 (2014) 66–72

Contents lists available at ScienceDirect

Journal of Neuroscience Methods

journal homepage: www.elsevier.com/locate/jneumeth

Basic Neuroscience

Diving deeper into Zebrafish development of social behavior:

Analyzing high resolution data
Christine Buske a,b,∗ , Robert Gerlai b
a
Papers/Springer SBM, London, UK (previously: University of Toronto, Department of Cell & Systems Biology)
b
University of Toronto Mississauga, Department of Psychology, Toronto, Canada

h i g h l i g h t s

• Zebrafish are a high throughput, cost effective vertebrate model.

• Behavioral data collection/analysis is time-consuming.
• The R programming language is a powerful tool in data analytics.
• R was used to analyze a large behavioral dataset.

a r t i c l e i n f o a b s t r a c t

Article history: Vertebrate model organisms have been utilized in high throughput screening but only with substantial
Received 2 May 2014 cost and human capital investment. The zebrafish is a vertebrate model species that is a promising and
Received in revised form 16 June 2014 cost effective candidate for efficient high throughput screening. Larval zebrafish have already been suc-
Accepted 16 June 2014
cessfully employed in this regard (Lessman, 2011), but adult zebrafish also show great promise. High
Available online 23 June 2014
throughput screening requires the use of a large number of subjects and collection of substantial amount
of data. Collection of data is only one of the demanding aspects of screening. However, in most screening
Keywords:
approaches that involve behavioral data the main bottleneck that slows throughput is the time consum-
Zebrafish
Behavior ing aspect of analysis of the collected data. Some automated analytical tools do exist, but often they
Methods only work for one subject at a time, eliminating the possibility of fully utilizing zebrafish as a screening
tool. This is a particularly important limitation for such complex phenotypes as social behavior. Testing
multiple fish at a time can reveal complex social interactions but it may also allow the identification of
outliers from a group of mutagenized or pharmacologically treated fish. Here, we describe a novel method
using a custom software tool developed within our laboratory, which enables tracking multiple fish, in
combination with a sophisticated analytical approach for summarizing and analyzing high resolution
behavioral data. This paper focuses on the latter, the analytic tool, which we have developed using the
R programming language and environment for statistical computing. We argue that combining sophis-
ticated data collection methods with appropriate analytical tools will propel zebrafish into the future of
neurobehavioral genetic research.
© 2014 Published by Elsevier B.V.

1. Introduction adult zebrafish have been becoming increasingly popular in

Larval zebrafish have been an excellent vertebrate model for
high throughput studies (Lessman, 2011), but adult zebrafish may
also hold substantial promise in this regard (Brennan, 2014). The
limiting factors in using adult zebrafish for high throughput stud-
ies lie not in maintenance or acquisition costs, but rather in time
consuming experimental procedures and analysis. Nevertheless
∗ Corresponding author. Tel.: +44 7867410264.
E-mail address: Christine.Buske@springer.com (C. Buske).
behavioral neuroscience as the identification of mutation or drug-
induced alterations in brain function may be best investigated with
behavioral test paradigms (Gerlai, 2010). While larval zebrafish
offer behavioral endpoints, adult zebrafish provide the advantage
of a far wider ranging behavioral repertoire (Norton and Bally-
Cuif, 2010). Behavior is particularly appropriate when modeling
human conditions for which abnormal behavior is a core symptom
(Gerlai, 2012; Brennan, 2014). Previous work has also indicated a
correlation between neurochemical changes and behavioral mat-
uration in zebrafish (Buske and Gerlai, 2012). From this leads the
argument that changes in behavior are accompanied by changes in

http://dx.doi.org/10.1016/j.jneumeth.2014.06.019
0165-0270/© 2014 Published by Elsevier B.V.
 C. Buske, R. Gerlai / Journal of Neuroscience Methods 234 (2014) 66–72
neurochemistry, and as we further develop our understanding of
how these differences interplay we may gain further understanding
of the brain and behavior.
Zebrafish offer a rich behavioral repertoire. Shoaling is one of
the behavioral outputs belonging to this (Buske and Gerlai, 2011a,b;
Miller and Gerlai, 2011). Several behaviors have been characterized
in zebrafish and are being further investigated within a behavioral
neuroscience context. These have been previously reviewed and
include reward, learning and memory, aggression, locomotion, anx-
iety, mating, and sleep (Norton and Bally-Cuif, 2010). Shoaling is a
complex behavior to study and quantify as there are a number of
protocols that mimic a shoaling situation, while quantifying shoal-
ing in an open field setting (where subjects are able to interact with
each other) has been difficult in the past (Buske and Gerlai, 2011a,b;
Gerlai, 2014).
While behavior offers insights into brain function, recording and
analysis of behavioral trials can be time consuming. Throughput
in behavioral screening may be successfully increased by running
several testing arenas in parallel, and cost of set up such systems is
decreasing as technology advances. This makes parallel behavioral
test systems a reality even for the average academic laboratory. The
bottleneck in such behavioral screening then becomes the proper
extraction and analysis of the acquired data. Several commercially
available tracking systems exist for testing a single subject at a time,
(e.g. Noldus’ Ethovision, and CleverSys). These systems are highly
sophisticated, but still present some trade-offs: while accurately
tracking a single fish in optimal conditions, these systems present
difficulty tracking very small fish in a large area (in larger tanks),
and require optimal light conditions. The additional disadvantage is
that most of these systems have difficulties with tracking multiple
fish, particularly shoals with more than four members. The final,
but certainly not most trivial limiting factor is the cost: With a sin-
gle license one could only analyze one trial at a time. This presents
a significant time constraint with regards to data extraction from
videotaped trials. Processing video files at a higher throughput rate
would require the purchase of multiple licenses and an equal num-
ber of computers. With tracking systems costing $5–15 thousand
per license, parallel processing (multiple licences) is not possible
for most academic laboratories.
Several commercial software applications exist that allow for
automated tracking of zebrafish, and quantification of several
behavioral outputs. As discussed above, these can be costly and the
cost can limit throughput. Having said that, these methods do offer
increasingly sophisticated means of measuring behavior: tracking
of animal behavior has been possible in a two dimensional plane for
some time (Noldus and Spink, 2001). More recently, various groups
have started applying 3D video behavioral tracking (Maaswinkel
et al., 2013). These methods have provided insight in various behav-
ioral outputs, particularly as a result of drug exposure (Cachat et al.,
2013; Maaswinkel et al., 2013).
When developing our in-house tools, first we consider the
requirements of the tracking system and the specifics of behavioral
analysis as they pertain to zebrafish. High throughput capabilities
would require a simple tracking system and reliable behavioral
paradigms (Blaser and Gerlai, 2006). The tracking system should
ideally be capable of tracking shoals of fish and tracking even under
sub-optimal conditions (e.g. when reflections cannot be avoided or
under sub-optimal light/background conditions). Preferably, such
systems should be capable of processing video-data in real time
and preferably with a small number of computers. For example,
under such conditions commercially available tracking systems fre-
quently mistake small particles, e.g. air bubble or debris, which
introduces substantial errors. The errors need to be corrected by the
experimenter, and this requires continuous monitoring and time.
Recent improvements in optics and computing power have
revolutionized the analysis of behavior. Just under 30 years ago,
67
tracking with one camera and one computer cost thousands of dol-
lars and was only able to generate a position for the animal once
per second. Now the cost for basic equipment is in the few hun-
dred dollar range and tracking occurs at 30 times per second with
a simple inexpensive camera (Dusenbery, 1985).
The rapid improvements in electronics and video equipment
have opened up the possibility of extracting large amounts of
data. Extracting positional data from a video recorded at 30 frames
per second results in 18,000 data points in a short 10-min video.
Processing and analysis that requires manual input would be pro-
hibitively time consuming for the number of trials required in the
average experiment. As technology improves, Big Data is becoming
the new bottle neck across many disciplines (Marx, 2013).
Shoaling behavior is complex, involving subjects leaving and
rejoining the shoal, and varying distances between shoal members.
Several research groups have attempted to describe and quan-
tify shoaling behavior using multiple zebrafish simultaneously, but
these approaches typically involve a method where the number of
individuals occupying an arbitrary area of space within the test-
ing tank is counted at different time intervals (Echevarria et al.,
2011), or by assigning a shoal cohesion ‘score’, a subjective judg-
ment made by the observer, at different time points during the
trial (Piato et al., 2011). In other studies, a single fish is exposed
to a shoal of conspecifics separated by a glass divider, and the
time spent within a preset compartment next to the stimulus is
measured (Savio et al., 2012). Aside from being time consuming,
these methods do not offer an objective or particularly informa-
tive description of shoal cohesion. For example, a group of fish may
be divided across two different arbitrary areas as defined by the
experimenter, but be physically very close to each other. In another
scenario, the group may be present within the same arbitrary area
of the testing arena, but be physically further apart. Shoal cohe-
sion would be rated lower in the first scenario than the second
and does not accurately represent how close the shoal truly is. In
cases where a shoal cohesion score is assigned to each time interval
sampled, substantial information is also lost, and there is the pos-
sibility of experimenter bias. Subtle differences in shoal cohesion
would be missed in such a method, even when assessed by multi-
ple raters and inter-rater reliability is high. Needless to say, these
methods do not allow for high throughput analysis of behavioral
trial, aside from not providing objective and precise measurements
of group behavior. Reproducibility of these methods across labo-
ratories is also of concern (Benjamini et al., 2010). Also notably,
methods where a single fish is tested in the presence of a stimulus
shoal (Savio et al., 2012) do not describe shoal behavior properly
as there is no possibility of interactive communication between
the subject and the stimulus. For example, even when live stim-
ulus fish are used, the experimental zebrafish will not be able to
sense the presence of the stimulus fish with their lateral line, as a
glass barrier is placed between the test subject and the stimulus
fish. Similarly, while larval zebrafish can be observed and tracked
in multiwell plates, each subject is isolated in its own well, and thus
the subjects do not interact (Cario et al., 2011). When measuring
shoaling in a group setting automated tracking methods have still
fallen short.
An earlier tracking program developed within our laboratory
(described in detail previously by Miller and Gerlai (2007) have
allowed for more objective and accurate measurement of inter-
individual distance, and other parameters of group behavior, in
zebrafish. This method has been successfully employed in previous
studies (Buske and Gerlai, 2011a,b). Notably, it relied on manually
identifying each individual in a shoal at each time interval sam-
pled, which required a human observer, and thus the method was
highly time consuming. Because the observer was only identifying
the location of the subject on a screen by clicking on it, and not mak-
ing any assessment on shoal cohesion, this method was arguably
68 C. Buske, R. Gerlai / Journal of Neuroscience Methods 234 (2014) 66–72
more objective than several prior manual scoring or rating meth-
ods. However, extracting high-resolution data from trials using this
method is prohobitively time consuming, and as a consequence it
is impossible to avoid loss of information unless a substantial time
investment is made.
More recently, a sophisticated yet simple tracking system was
developed in our laboratory, internally named ‘Real Fish Tracker’. It
was developed by James McCrae, a computer science PhD Candi-
date at the University of Toronto. This tracking software is able to
track multiple subjects within the same environment, and records
precise location data (X-Y coordinates) for each fish at the frame
rate of the video being sampled. This translates to a sampling rate
of 29 times per second for videos created with conventional dig-
ital cameras. The program not only records location data at 29×
per second, but it does so in real time on computers with average
processing speeds. In addition, multiple sessions can run simulta-
neously on the same computer, allowing for tracking of up to five
different trials in unison on a laptop computer with a modest pro-
cessor and memory card. It should be noted that when running this
many instances of the program, tracking does not proceed in real
time, but some time savings is gained from only calibrating a few
trials at once every half an hour instead of calibrating one file every
few minutes.
The program supports a range of conventional video formats,
and tracks the fish by comparing the frame being sampled with
an average image computed over the previous several frames in
the video. The difference in images allows the program to identify
the change in pixels, i.e. the location of the fish, and assigns x-y
coordinates to each subject. The algorithms take into account the
previously known position of the subject, the current image and
the average images for previous frames. This allows for a highly
reliable determination of the position of each subject, and mini-
mizes instances where a subject might be ‘lost’ by the program due,
for example, to two subjects crossing into the path of each other.
While minimizing these instances, they do occur. The software, as
with any other application, does not guarantee consistent identi-
fication of the same fish. This would be of concern in paradigms
where an individual mutant or differentially treated fish may be
exposed to shoal of control fish. The experiments discussed within
this article are focused on characterizing dynamics of movement
of the shoal and as such do not require individual labeling of each
shoal member, an additional goal that will require future software
development.
The current software program provides high-resolution posi-
tional data for each of the subjects in the trial. However, it does
not provide any quantification of particular (established) behav-
ioral endpoints, such as inter-individual distance, nearest neighbor
distance, distance from the closest wall or center, distance from a
corner, etc. As described before, some sophisticated software pack-
ages exist that can analyze select behavioral outputs in zebrafish
(Ahmad et al., 2012), but for many research groups the cost of acqui-
sition of these packages can still be prohibitive. In addition, these
packages offer a set of behavioral outputs that can be quantified,
but beyond these predetermined behavioral measures, they do not
offer more flexibility. The current method was deployed in a dif-
ferent study validating its reliability for the endpoints measured;
comparable results were obtained in the description of shoaling
behavior with the currently described method (Buske and Gerlai,
2012; Mahabir et al., 2013), as with older studies using previously
discussed methods (Buske and Gerlai, 2011a,b).
Several other tracking programs exist that have been either cus-
tom developed or open source (Aguiar et al., 2007; Wolfer et al.,
2001; Kane et al., 2004). The community faces the challenge of
frequently reinventing the wheel, with several tracking programs
being developed with sometimes overlapping features. Most of
these programs produce data files consisting of coordinate data in
a xy field for the subject being tracked. At this stage, the analysis of
the data provides the second major challenge.
Using the raw data output from the application developed in-
house affords this flexibility. We have developed a quantification
module based on the R-environment for statistical computing that
allows us to extract numerous behavioral endpoints from the raw
Data-output, and in a highly flexible, user-specific manner. In our
behavioral experiments, we recorded the behavior of our fish for
8 min. Each 8-min trial resulted in data files consisting of 13,920
rows of data (sampled 29× per second). More complex behav-
ioral studies, e.g. those requiring following the subject’s behavior
across extended periods of time, generate even larger data matri-
ces. A longitudinal study recently completed, assessing the effects
of embryonic ethanol exposure over the course of development,
generated 1300 data files (13 age points, n = 20 for five treatment
groups). This resulted in a total of 18 million data points (Buske and
Gerlai, 2011a,b).
Processing and computation of many hundreds, or thousands
of data files corresponding to an equal number of trials require a
sophisticated approach. The R programming language and envi-
ronment for statistical computing is particularly suitable for this
purpose (Venables and Smith, 2012). R can be regarded as an
implementation of the S language which was developed at Bell Lab-
oratories by Rick Becker, John Chambers and Allan Wilks, and also
forms the basis of the S-Plus systems. It is an effective program-
ming language that facilitates data manipulation, calculation, and
graphics.
The software suite is referred to as an ‘environment’, as it is a
system for developing methods of interactive data analysis, rather
than a typical statistical package or data analysis software. R pro-
vides the researcher with flexibility in designing analytical tools
suitable for very specific data sets or goals unique to a particular
project.
By creating programs in R for the data manipulation and analysis
of high resolution positional data as acquired with the Real Fish
Tracker program it is possible to quickly and efficiently process and
analyze hundreds of output files in a very short time frame, and
with minimal interference by the experimenter.
2. Data processing and analysis
Similar to several other tracking programs, The Real Fish Tracker
produces individual .txt format output files with high resolution
positional time series data for each individual tracked within the
same arena. Video files are sampled at a rate of 29× per second,
creating data files of thousands of rows with data.
Each data file corresponds to a single trial, and from the posi-
tional data recorded various different behavioral measures can be
computed in a user-defined and highly flexible manner. For our
own purposes we decided to compute the following behavioral
measures: inter-individual distance, nearest neighbor distance, dis-
tance from the closest wall or center, distance from a corner,
distance traveled, time spent in a perimeter or particular zone, etc.
As only positional data are provided, further processing is both a
must and a great advantage providing flexibility. The 29 hz samp-
ling rate generates many time points and the average of these
(the temporal mean) is calculated for the above measures for user
defined intervals. In addition, but also importantly, the variance
of the time point data is also calculated. It represents the within
individual temporal variance of the behavior.
The xy coordinates are in screen units and need to be con-
verted to an experimenter defined distance unit, e.g. cm or body
lengths. The latter is commonly used in animal, and particularly
fish, research (Partridge, 1981). Subsequently, a new data frame
can be constructed with the means of these time intervals or trials
 C. Buske, R. Gerlai / Journal of Neuroscience Methods 234 (2014) 66–72
for statistical analysis. Processing each data file individually is pro-
hibitively time consuming and an automated procedure for high
throughput processing is a necessity. The development of an ana-
lytical tool in the R programming language, to complement the
tracking program solves this issue and is described below.
3. Requirements
Batch automation of all script files applied to a selection of
data files has been accomplished using the args package (Piipari,
2010). The R environment runs on both Macintosh and Widows
based operating systems. Basic knowledge of data processing in R
is required for basic analysis.
3.1. Data processing workflow
Raw data files are in .txt format and contain the following vari-
ables:
• Tracking area (coordinate points)
• Calibration rulers x and y (coordinate points), and length (in cm
or other measure defined by the user).
• Seconds
• FishX coordinates for each subject
• FixhY coordinates for each subject
• Confidence (a confidence measure for detecting the subject in the
trial)
• Ruler 1 and Ruler 2 calibrated positions for each subject.
R scripts are written to process and analyze the data obtained
in an automated manner.
3.2. Data pre-processing
The first R script applied to the data is to reshape the data for
further analysis. It is possible to perform ANOVAs and other statis-
tical analyses in R using data in a long format. The data obtained
from the tracking program is already in the long format, but con-
tains calibration information in the first six rows of the matrix. The
original data sets can be simplified first for easier processing and
behavioral quantification.
The calibration information is transferred to a new column in
the data matrix. For functions to be applied across the entire data
matrix, the R script written requires the matrix to be of equal width
and length. All pertinent information retained in the first rows of
the output file is therefore reshaped to fit this requirement. In the
interest of processing speed, non-essential information is removed
from the new data matrix. All original data files are preserved, and
the script outputs a newly reshaped data matrix.
First, the calibration information is called from each trial:
tanksize <- as.numeric(as.character(trial[6,5]))
xa <- as.numeric(as.character(trial[4,1]))
xb <- as.numeric(as.character(trial[4,3]))
ya <- as.numeric(as.character(trial[4,2]))
yb <-as.numeric(as.character(trial[4,4]))
xa, xb, ya, and yb are standard arguments used across the script
library to denote each of the corners of the tank (Fig. 1). In the data
file the calibration information is contained within a specific loca-
tion that is automatically queried using these arguments and the
locations of which are denoted between the square brackets. By
performing this step, it normalizes each of the different xy coor-
dinates for the arena calibration so that despite each trial having
a slightly different xy coordinate for the corners of the arena, the
same script can be applied to all data files (Fig. 1).
69
Fig. 1. Calibration of the testing arena. All calibration information, which is unique
for each trial, is captured by the R scripts and tranformed to a numeric object. This
allows for applicability of each script across all trials, despite different calibration
information.
Subsequently, the calibration information is skipped and the
required columns from the data matrix are selected, and the new
matrix is saved.
3.3. Calculation of inter-individual distances
After the pre-processing step, a new R script is applied to all
matrices generated from the first step described above. In this
script, formula to calculate the distance between each focal fish and
each other fish present in the arena are applied across the entire
data table, generating 10 new columns of distance data for each of
the ten focal fish in a trial.
The formula used follows from the Pythagorean theorem:
f1f2d2 <- ((trial.sub$f2rx-trial.sub$f1rx)2̂ + (trial.sub$f2ry-
trial.sub$f1ry)2̂)
f1f2d <- f1f2d2ˆ(1/2)
The code above describes the distance between fish 1 and fish
2. The same is repeated within the script for each other focal fish
for all unique distance combinations.
Distances are taken from the calibrated ruler information pro-
vided in the data matrix, and are thus set in the measure for which
the arena was calibrated. For each of our experiments, calibration
was done in centimeters, but naturally any distance measure may
be used as desired by the user.
All vectors containing the distances of one focal fish to one other
focal fish are combined into a new object called indiv.dist, and the
mean of each row is calculated (mean of time point inter-individual
distance data for each time interval sampled).
indiv.dist <- cbind(f1f2d, f1f3d, f1f4d, f1f5d, f1f6d, f1f7d, f1f8d,
f1f9d, f1f10d, f2f3d, f2f4d, f2f5d, f2f6d, f2f7d, f2f8d, f2f9d, f2f10d,
f3f4d, f3f5d, f3f6d, f3f7d, f3f8d, f3f9d, f3f10d, f4f5d, f4f6d, f4f7d,
f4f8d, f4f9d, f4f10d, f5f6d, f5f7d, f5f8d, f5f9d, f5f10d, f6f7d, f6f8d,
f6f9d, f6f10d, f7f8d, f7f9d, f7f10d, f8f9d, f8f10d, f9f10d)
average.dist <- rowMeans(indiv.dist, na.rm=FALSE, dims=1)
The new vector with the mean inter individual distance is added
to the individual distance vectors. Subsequently, this data frame is
70 C. Buske, R. Gerlai / Journal of Neuroscience Methods 234 (2014) 66–72
merged with the original data matrix containing the XY coordinate
data. Using a small matrix containing independent factor informa-
tion, for example in our case the age, strain, treatment, and body
length of the subjects is added to the data matrix.
From each of the inter-individual distances the minimum dis-
tance is called and a new vector is created, representing the
minimum-neighbor distance. This is different from the nearest
neighbor distance in the sense that nearest neighbor distance rep-
resents the average smallest inter individual distance among all
focal fish. The minimum-neighbor distance is the smallest value
for all focal fish. Both measures are computed.
Subsequently, the distance for each focal fish from each of the
walls of the arena is extracted using the raw xy coordinate data
matrix. At each time point measured, four new vectors are calcu-
lated for each of the subjects in the arena. Each vector represents
the distance of the focal fish from one of the walls. The minimum
of the row is taken for each fish, and summarized in a new vector.
The example code for fish 1 is as follows:
f1.dist.xa <- (trial$f1.x-xa)
f1.dist.xb <- (xb-trial$f1.x)
f1.dist.ya <- (trial$f1.y-ya)
f1.dist.yb <- (yb - trial$f1.y)
f1.wall <- cbind(f1.dist.xa, f1.dist.xb, f1.dist.ya, f1.dist.yb)
f1.min <- apply(f1.wall, 1, min)
After applying the script to a data file, entering the object name
f1.min, and hitting enter returns the time series for the distance to
the closest wall for fish 1.
Each vector is created using a collection of formulae and saved
in R as an object. The object can be called from the R script for each
individual data file as long as the entire script is run first.
3.4. Extracting data
The R library created consists of 20 separate scripts customized
for either single subject data or multi-subject data.
Data from each trial can be summarized and the object of
interest (measure, e.g. inter-individual distance, nearest neighbor
distance, etc.) can be called for either the complete trial or for time
windows within each trial. For each trial these values can be col-
lected into one data matrix for statistical analysis across treatment
groups.
We created a customized script containing commands for each
of the objects to be called for from the library. Then, this script can
be applied over an unlimited number of data files to quickly and effi-
ciently process and summarize data for all trials in an experiment.
Within the same script, the possibility of exporting a file contain-
ing all statistical analyses can be included. This automates the data
processing and analysis steps of a research project. By allowing the
program to export the summary tables for each of the measures we
include the possibility for the experimenter to perform additional
statistical analyses, in addition to being able to export to a standard
statistical analysis as desired by the user (Figs. 2 and 3).
4. Discussion & future directions
High throughput screening is typically thought of in the
context of simple observable characteristics in small model orga-
nisms (Lessman, 2011) but with increasingly powerful behavioral
paradigms and methodologies now screening is becoming feasi-
ble even for complex behavioral phenotypes. Adult zebrafish are
a small and highly social species (Buske and Gerlai, 2011a,b) and
the above methods now suggest that high throughput screens
for complex behaviors including numerous parameters of group
Fig. 2. Explanation of the distance to the closest wall measurement: For each focal
fish (red) the distances to each of the four walls of the arena is calculated based on
the individual’s location on the XY coordinate grid. Each of these wall distances is
saved to a separate vector. For these four vectors, the minimum value is extracted
into a fifth vector and used as the minimum wall distance. The original four vectors
are discarded after the calculation is completed across all time points sampled. (For
interpretation of the references to color in this figure legend, the reader is referred
to the web version of this article.)
forming (shoaling) and exploratory behavior are becoming a reality.
Hardware and software solutions for screening with zebrafish larva
have become commercially available and are usually performed
in 96 well plates (Redfern et al., 2008; Giacomotto and Ségalat,
2010; Schnorr et al., 2012). Adult zebrafish provide additional ben-
efits not present yet in larvae; e.g. a more sophisticated behavioral
repertoire (Norton and Bally-Cuif, 2010) such as complex forms of
social behavior including shoaling (Buske and Gerlai, 2011a,b). We
have previously shown that shoaling can be disrupted at the adult
stage by a single embryonic insult with a teratogen such as ethanol
(Fernandes and Gerlai, 2009; Buske and Gerlai, 2011a,b) Behavioral
tests can provide insights into brain function and may for example
help in the analysis of the effects of early embryonic teratogenic
insults (Schnorr et al., 2012). Utilizing adult zebrafish as a model to
study altered brain function in a vertebrate is particularly attractive.
However, when attempting to identify mutants or efficacious drugs
from a library of compounds, high throughput methods are neces-
sary (Zon and Peterson, 2005). Housing and testing adult zebrafish
is economical and thus the practical and financial bottleneck in high
throughput studies lies in the extraction and analysis of data.
Previously, manual tracking techniques only partially auto-
mated the process of obtaining reliable and quantifiable data from
behavioral recordings (Miller and Gerlai, 2007). The time invest-
ment required not only for the tracking aspect, but also the manual
processing of the data in a spreadsheet program such as Microsoft’s
Excel is prohibitively time consuming with Big Data. Human error
and bias are also of some concern with such approaches. Even
when using automated tracking software, the raw data gener-
ated still requires substantial preparation and processing prior to
downstream analyses. Furthermore, commercial automated track-
ing software is expensive, and acquisition of sufficient licenses to
run a high throughput operation is prohibitive for the average lab-
oratory. These systems have so far proven inefficient or unable to
track multiple subjects at the same time in the same arena, severely
limiting zebrafish behavioral research. We address each of these
 C. Buske, R. Gerlai / Journal of Neuroscience Methods 234 (2014) 66–72 71

Fig. 3. Explanation of minimum nearest neighbor (MNND) vs. nearest neighbor distance (NND). Each circle represents an individual (i.e. a fish). The Nearest Neighbor Distance
(NND) of the focal (red) individual is the distance between it and the closest other individual, represented by the red arrow. The NND is calculated as mean of all nearest
neighbor distances for each focal fish (B). The Minimum nearest neighbor distance only measures the minimum of all nearest neighbor distances for a sampled time point,
so only the smallest nearest neighbor distance is calculated as shown in (A). (For interpretation of the references to color in this figure legend, the reader is referred to the
web version of this article.)

limitations with a novel approach consisting of a custom tracking
application and analytical tool developed in our laboratory.
The tracking system described generates multi-subject track
data in an xy coordinate format. The raw data generated by The
Real Fish Tracker can be now processed and analyzed so that often
complex behavioral measures are extracted and quantified in an
automated manner using the R language (Gentleman et al., 2004).
Using a command line interface (such as Terminal app on Macintosh
computers, and Windows Command Processor on PCs), the scripts
developed for the desired behavioral measures can be applied to
an unlimited number of data files within a project folder. Just as an
example, it is possible to process and analyze 1300 individual data
files (containing a total of 18 million rows and 28 columns of data)
within approximately 3 h using a Macintosh computer running OSX
10.8, with a 2.6 GHz Intel core i7 processor and 16 GB 1600 MHz
DDR3 of memory. It would take several weeks or even months for
an individual to process the same amount of data and perform all
calculations, even when assisted with e.g. an Excel Macro for some
of the tasks.
At this time the library of R scripts we have developed for these
analytical purposes exists as a separate set of script files. Moving
forward, it will be necessary to rewrite aspects of these for effi-
ciency and processing speed, and combine them together in the
R package. This package can then easily be distributed to anyone
who is interested through the Comprehensive R Archive Network
(CRAN), and through any other means deemed appropriate (e.g.
potentially a dedicated website where both the tracking software
and the analytical package can be offered.)
At the time of writing the CRAN repository features 3971 avail-
able packages. Each of these has been contributed by a member of
the community, and made available to anyone else interested in
using the analytical tools contained within them. This illustrates
the breadth and activity of the R programming community, and
the R user community as a whole. It also emphasizes that steps in
the analytics process do not need to be recoded, and the prover-
bial reinventing the wheel does not need to occur, when another
researcher has contributed an appropriate resource already.
R is open source software, making it accessible to virtually any-
one, anywhere. The number of R users today is estimated to be in
the millions, across a variety of different disciplines and industries.
Although R is a command line driven system, development of an
R package containing many of the behavioral outputs discussed in
this study will minimize the R-specific experience and knowledge
required to operate the package, while maintaining the flexibility of
the R environment for more experienced users. The popularity for
the R language has grown dramatically in recent years. The R lan-
guage is an appropriate choice from a methodological perspective.
In addition, with a growing number of users and the open source
nature of the platform, there are many opportunities for innovation
and collaboration.
The behavioral outputs discussed in this article may be viewed
as examples and starting points of a vast array of future possibilities.
Any behavior that can be quantified using xy positional data can be
extracted using these methods. One does not need to rely on pre-
set parameters that can be quantified, and instead the researchers
has free reign of how in depth the analysis proceeds.
In addition to benefits of analytical flexibility, these methods
are attractive to the large number of laboratories that are not able
to afford costly commercially available tracking systems. While the
zebrafish is growing in popularity in behavioral neuroscience (Sison
et al., 2006; Gerlai, 2010), the aspects involving data acquisition
and processing in behavioral studies do not match the low cost and
high throughput capabilities typical of research conducted with
this species. Our tracking software, combined with basic R skills
for pre-made R scripts can make high powered analytics accessible
at minimal cost. This eliminates one of the primary bottlenecks in
high throughput screening of adult zebrafish, and high throughput
behavioral testing in general. These types of software applications
72 C. Buske, R. Gerlai / Journal of Neuroscience Methods 234 (2014) 66–72
and analytical tools are not limited to zebrafish, they can be applied
with other model organisms as well. Therefore, we hope that our
methods will gain popularity across a broad spectrum of Big Data
applications and will make analysis of results easier and more effi-
cient.
References
Aguiar P, Medonça L, Galhardo V. OpenControl. A free opensource software for
video tracking and automated control of behavioral mazes. J Neurosci Methods
2007;166(October (1)):66–72.
Ahmad F, Noldus L, Tegelenbosch R, Richardson M. Zebrafish embryos and larvae in
behavioural assays. Behavior 2012;149(January (10–12)):1241–81.
Benjamini Y, Lipkind D, Horev G, Fonio E, Kafkafi N, Golani I. Ten ways to improve
the quality of descriptions of whole-animal movement. Neurosci Behav Rev
2010;34(July (8)):1351–65.
Blaser R, Gerlai R. Behavioral phenotyping in zebrafish: comparison of three behav-
ioral quantification methods. Behav Res Methods 2006;38(August (3)):456–69.
Brennan CH. Zebrafish behavioural assays of translational relevance for the study of
psychiatric disease. IEEE 2014;22(1).
Buske C, Gerlai R. Early embryonic ethanol exposure impairs shoaling and the
dopaminergic and serotoninergic systems in adult zebrafish. Neurotoxicol Ter-
atol 2011a;33(November (6)):698–707.
Buske C, Gerlai R. Shoaling develops with age in Zebrafish (Danio rerio). Prog Neuro-
Psychopharmacol Biol Psychiatr 2011b;35(August (6)):1409–15.
Buske C, Gerlai R. Maturation of shoaling behavior is accompanied by changes
in the dopaminergic and serotoninergic systems in zebrafish. Dev Psychobiol
2012;54(January (1)):28–35, Wiley Subscription Services, Inc., A Wiley Com-
pany.
Cachat J, Kyzar EJ, Collins C, Gaikwad S, Green J, Roth A, et al. Unique and potent
effects of acute ibogaine on zebrafish: the developing utility of novel aquatic
models for hallucinogenic drug research. Behav Brain Res 2013;236(January
(1)):258–69.
Cario CL, Farrell TC, Milanese C, Burton EA. Automated measurement of zebrafish
larval movement. J Physiol Physiol Soc 2011;589(August (15)):3703–8.
Dusenbery D. Using a microcomputer and video camera to simultaneously track 25
animals. Comput Biol Med 1985;15(4).
Echevarria DJ, Buske C, Toms CN, Jouandot DJ. A novel test battery to assess
drug-induced changes in zebrafish social behavior. zebrafish neurobehavioral
protocols. Totowa, NJ: Humana Press; 2011. p. 109–24.
Fernandes Y, Gerlai R. Long-term behavioral changes in response to early devel-
opmental exposure to ethanol in Zebrafish. Alcohol Clin Exp Res 2009;33(4):
601–9.
Gentleman R, Carey V, Bates D, Bolstad B, Dettling M, Dudoit S, et al. Bioconduc-
tor: open software development for computational biology and bioinformatics.
Genom Biol 2004;5(10):R80.
Gerlai R. High-throughput behavioral screens: the first step towards finding genes
involved in vertebrate brain function using zebrafish. Molecules 2010;15(April
(4)):2609–22.
Gerlai R. Using zebrafish to unravel the genetics of complex brain disorders. Behav
Neurogen 2012;12:3–24.
Gerlai R. Social behavior of zebrafish: from synthetic images to biological mecha-
nisms of shoaling. J Neurosci Methods 2014.
Giacomotto J, Ségalat L. High-throughput screening and small animal models, where
are we? Br J Pharmacol 2010;160(March (2)):204–16.
Kane A, Salierno J, Gipson G, Molteno T, Hunter C. A video-based movement
analysis system to quantify behavioral stress responses of fish. Water Res
2004;38(November (18)):3993–4001.
Lessman C. The developing zebrafish (Danio rerio): a vertebrate model for high-
throughput screening of chemical libraries. Birth Defects Res Part C: Embryo
Today: Rev 2011;93(3):268–80.
Maaswinkel H, Le X, He L, Zhu L, Weng W. Dissociating the effects of habitua-
tion, black walls, buspirone and ethanol on anxiety-like behavioral responses
in shoaling zebrafish. A 3D approach to social behavior. Pharmacol Biochem
Behav 2013;July (108):16–27.
Mahabir S, Chatterjee D, Buske C, Gerlai R. Maturation of shoaling in two
zebrafish strains: a behavioral and neurochemical analysis. Behav Brain Res
2013;247(June):1–8.
Marx V. Biology: the big challenges of big data. Nat Technol 2013;498(June
(7453)):255–60.
Miller N, Gerlai R. Quantification of shoaling behaviour in zebrafish (Danio rerio).
Behav Brain Res 2007;184(December (2)):157–66.
Miller N, Gerlai R. Shoaling in zebrafish: what we don’t know. IEEE 2011;22(1).
Noldus L, Spink AJ, Tegelenbosch RAJ. EthoVision. A versatile video tracking system
for automation of behavioral experiments. Behav Res Methods Instrum Comput
2001;33(August (3)):398–414, Springer-Verlag.
Norton W, Bally-Cuif L. Adult zebrafish as a model organism for behavioural genetics.
BMC Neurosci 2010;11:90.
Partridge B. Internal dynamics and the interrelations of fish in schools. J Comp
Physiol 1981;144(3):313–25.
Piato A, Rosemberg D, Capiotti K, Siebel A, Herrmann A, Ghisleni G, et al. Acute
restraint stress in zebrafish: behavioral parameters and purinergic signaling.
Neurochem Res 2011;36(October (10)):1876–86.
Piipari M. Inference and classification of eukaryotic cis-regulatory motifs;
2010, Available from ftp://anonymous@ftp.sanger.ac.uk/pub4/theses/mp4/
abstract.pdf
Redfern W, Waldron G, Winter M, Butler P, Holbrook M, Wallis R, et al. Zebrafish
assays as early safety pharmacology screens: paradigm shift or red herring? J
Pharmacol Toxicol Methods 2008;58(September (2)):110–7.
Savio L, Vuaden F, Piato A, Bonan C, Wyse A. Behavioral changes induced by
long-term proline exposure are reversed by antipsychotics in zebrafish. Prog
Nueropsychoparmacol Biol Psychiatr 2012;36(March (2)):258–63.
Schnorr S, Steenbergen P, Richardson M, Champagne D. Measuring thigmotaxis in
larval zebrafish. Behav Brain Res 2012;228(March (2)):367–74.
Sison M, Cawker J, Buske C, Gerlai R. Fishing for genes influencing vertebrate behav-
ior: zebrafish making headway. Lab Anim (NY) 2006;35(May (5)):33–9.
Venables W, Smith D. R Core Team. An Introduction to R: Notes on R, a Programming
Environment for Data Analysis and Graphics; 2012 http://cran.r-project.org/
doc/manuals/R-intro.pdf
Wolfer D, Madani R, Valenti P, Lipp H. Extended analysis of path data from mutant
mice using the public domain software Wintrack. Physiol Behav 2001;73(August
(5)):745–53.
Zon L, Peterson R. In vivo drug discovery in the zebrafish. Nat Rev Drug Discov
2005;4(January (1)):35–44.