Blackiston D et al. (2010), A second-generation device for automated traini...

XB-ART-42543

PLoS One 2010 Dec 17;512:e14370. doi: 10.1371/journal.pone.0014370.

Show Gene links Show Anatomy links

A second-generation device for automated training and quantitative behavior analyses of molecularly-tractable model organisms.

Blackiston D , Shomrat T , Nicolas CL , Granata C , Levin M .

???displayArticle.abstract???
A deep understanding of cognitive processes requires functional, quantitative analyses of the steps leading from genetics and the development of nervous system structure to behavior. Molecularly-tractable model systems such as Xenopus laevis and planaria offer an unprecedented opportunity to dissect the mechanisms determining the complex structure of the brain and CNS. A standardized platform that facilitated quantitative analysis of behavior would make a significant impact on evolutionary ethology, neuropharmacology, and cognitive science. While some animal tracking systems exist, the available systems do not allow automated training (feedback to individual subjects in real time, which is necessary for operant conditioning assays). The lack of standardization in the field, and the numerous technical challenges that face the development of a versatile system with the necessary capabilities, comprise a significant barrier keeping molecular developmental biology labs from integrating behavior analysis endpoints into their pharmacological and genetic perturbations. Here we report the development of a second-generation system that is a highly flexible, powerful machine vision and environmental control platform. In order to enable multidisciplinary studies aimed at understanding the roles of genes in brain function and behavior, and aid other laboratories that do not have the facilities to undergo complex engineering development, we describe the device and the problems that it overcomes. We also present sample data using frog tadpoles and flatworms to illustrate its use. Having solved significant engineering challenges in its construction, the resulting design is a relatively inexpensive instrument of wide relevance for several fields, and will accelerate interdisciplinary discovery in pharmacology, neurobiology, regenerative medicine, and cognitive science.

???displayArticle.pubmedLink??? 21179424
???displayArticle.pmcLink??? PMC3003703
???displayArticle.link??? PLoS One
???displayArticle.grants??? [+]

Species referenced: Xenopus laevis
Genes referenced: acta4 ddx59 fry mmut pc.1 yes1

???attribute.lit??? ???displayArticles.show???

	Figure 1. Schematic diagram and physical construction of the automated behavior device.The system consists of a bank of individual cells connected to controllers, with the whole device running an embedded Linux system and communicating with a master PC computer via Ethernet networking. Panel (A) shows the top-level system architecture. The components are as follows. Training Apparatus Controller Gateway Device (TACGWD) and MNIC/MCM Device: The TACGWD acts as the logical TCP/IP interface between the host PC, MV Cameras, and all excitation related electronics. (The PoE Switch is the physical layer interface). Machine Vision Cameras (MV Cameras): The Machine Vision (MV) Cameras provide image acquisition and image processing functionality to determine the organism coordinates, size and other parameters related to the organism geometry. Channel Control Modules (CCM): The CCMs provide demultiplexing for excitation related communications between the TACGWD/MNIC and the ECMS. (I.E., each CCM interprets addressed multi-drop network commands for it's corresponding ECM channel). Excitation Control Module (ECM): The ECMs execute excitation-related commands from the CCM and provide control signaling to SCM and ICM hardware accordingly. The SCM and ICM contain no MCUs, so all firmware functionality resides in the ECM. Shock Control Module (SCM): The SCM provides a voltage controlled current source and 6 half-H bridge networks. Together these circuits generate a rotating AC current to drive the 6 electrodes in the Electrode Holder Assembly. Illumination Control Module (ICM): The ICM provides 4 quadrants of blue Negative Reinforcement (NR) Illumination and 4 quadrants of red Back-Ground (BG) illumination. (B) The general design includes an illumination head with four quadrants, containing blue and red LED's in each, that illuminates a petri-dish from above. Within each petri dish is a 12-sided electrode holder, hexagonal in design, that can deliver electric shock to the dish based on organism behavior. Below the dish is a red bypass filter (to remove all light except for red, simplifying background subtraction) and an individual camera which records video for each channel. Video feeds and electrical connections from the illumination head and electrode holder run behind the physical device and are processed/controlled by a separate electronics rack, which is in turn connected to a PC for control by the user. (C) The physical device is composed of 4 separate banks each containing 3 channels, for a total of 12 testing environments, and is raised off the floor for ease of access. (D) A close up of an open bank of channels reveals the illumination head with dividers, electrode holders, petri dishes. The connection running to the electronics rack can be seen in the background. The basic workflow is shown in (E). After loading the animals and setting up trial parameters on the master PC, the device runs (in parallel, for each dish independently) a cycle consisting of altering dish conditions (if needed), ascertaining coordinates of centroid of each animal, determining which animals' dish conditions are to be changed based on the trial type (e.g., shock applied as punishment, or lights turned off as reward), and writing current state data to log file. At the end, special scripts process all of the data and produce numerous statistics characterizing the behavior of each channel's subject.
	Figure 2. Example of the front end graphical user interface (GUI).Through this interface, the user defines the trial to be performed. Observation-only trials (useful to characterize behavior of mutants or pharmacologically-altered individuals) can be run simply by leaving out any shock or light feedback to the animals. In this example, the GUI is programmed for a planarian phototaxis trial as describe in Fig. 5. Notable features of this GUI include: (1) e-mail notification, both after a successful trial ends and immediately if a serious problem occurs during the trial; (2) yoked control mode can be indicated for half of the device chambers (the 6 even or odd channels); (3) light punishment can be associated with a specific position of the animal -4 quadrants can be set to yes (Y) or no (N) with specific distances from the edge or center or with general movement (speed), and the electric shock conditions can be set similarly (independently); (4) blue light illumination in different intensities can be set as background or punishment for each one of the quadrants; and (5) data quality threshold filter can detect abnormal animal movements like jumps of a large distance in a short time that may indicate a tracking problem. The threshold setting 0â1 will determine when punishment should be stopped in order to prevent false training.
	Figure 3. Application of light and electric shock to a Petri dish with aqueous medium.(A) A divider extends vertically from the LED assembly (without touching the water in the dish) and provides for excellent separation between light and dark quadrants (minimal light leakage). (A') The electrode assembly is composed of a round insert composed of white delrin, 6 iridium oxide-coated titanium electrodes, and control electronics. The width and height of the insert were designed for a snug fit with the walls and base of the petri-dishes used in experiments, to ensure planarians could not exit the testing environment from below. When shocks are issued to a given dish holder, two adjacent posts send and receive shock from the two opposite posts at an AC frequency specified by the user. Every 8 ms the source and sink electrodes rotate to the right, with a complete rotation occurring in 48 ms (the shortest shock that can be delivered). To ensure that animals received a nearly identical shock regardless of their position or orientation, finite element analysis was used to model the J field. While there are six physical electrodes, only 4 poles are shown in the model analyses (panels BâD) because only four are active during any given shock pulse. (B) A conventional 2-electrode design has a highly anisotropic field density, with hot spots near the electrodes and a defined polarity that will affect animals differently depending on their orientation with respect to the positive and negative poles, and the line connecting them. (C) A six-electrode design does better, but exhibits some dead spots as well as hot spots around the edge. (D) Much better homogeneity is obtained by using a 6-electrode design in which the electrodes take turns being the positive and negative pole. In this scheme, 82% of the dish area has a current density within 30% of the mean.
	Figure 4. Occupancy maps generated from Xenopus tadpole behavioral data to evaluate overall positional trends during trials.(A) Different quadrants of the dish can be set with varying light conditions (color and brightness), for instance half of the dish in red light and the other half in blue light. (B) Curiosity plots demonstrate overall preference for both light color and dish location (such as the edge or center of the dish), as illustrated by trials showing that tadpoles have no preference for either red or blue light. (C) When tadpoles receive a 1.4 mA electric shock under blue light, individuals now spend almost all of their time under the red illuminated half of the dish. Higher intensity of color (red) in heat plot indicates more time spent in that location.
	Figure 5. Comparison of planarian exploration behavior and light preference between D. japonica and S. mediterranea.During the trial one half of the training environment was illuminated with blue light and the other half with red. After one hour, illumination quadrants were exchanged so that the red half was now blue and vice versa (see arrow AâD). (A,B) Preference to light, the histograms summarize the percentage of time spent by the overall worms from each species (S. mediterranea in A and D. japonica in B) stayed in the red or blue illuminated half. Each category bar represents an average of 10 minutes, except for the first category (see arrow) that indicates the initial position of the worms just before the start of the trial, when all dish quadrants were illuminated by red light. Both species preferred the red light and were located in this half by the end of the exploratory phase (nâ=â23/24 D. japonica, nâ=â23/24 and S. mediterranea respectively); afterwards, they quit moving and settled down. Reversing the color light in each half induced just some of the subjects (13/24 of D. japonica but only 2/24 of S. mediterranea) to move into the red illuminated half. (C) Comparing movement rates and (D) explored area, both species demonstrated an exploratory phase, after which movement rates and exploration dropped to low levels. D. japonica (red triangles) showed significantly greater average speed and area explored than did S. mediterranea (black triangles). (E for S. mediterranea, F for D. japonica) Occupancy maps enable the ready evaluation of overall positional trends during trials. The plots generated from the behavioral data of the most active worm from each species during the first hour of the trial are shown. The red illuminated half is up in E and down in F, both indicating a preference for red over blue light as well as for the edge of the dish.
	Figure 6. Comparison of color preference and movement rates between 14-day-old Xenopus laevis tadpoles and 28-day-old Danio rerio fry.Tadpoles and fry were placed in the device with one half of each testing environment illuminated with red light and the other half illuminated with blue light. Images were recorded at 10 frames per second to accurately measure the movement rates of each organism. (A) Over a 30 minute initial preference trial, Xenopus tadpoles showed no preference for either blue or red light, while Danio fry spent more of their time under blue light. (B) Comparing movement rates, both organisms demonstrated a 15â20 minute exploratory phase, after which movement rates became stable, with zebrafish showing a greater average speed than Xenopus. Nâ=â9, 12 for Danio and Xenopus respectively, error bars indicate Â±1 SEM.
	Figure 7. Simple learning trial with Xenopus tadpoles.(A) Tadpoles 14 days of age were placed into the device individually, with half of the dish illuminated with low intensity red light and the opposite half illuminated with high intensity blue light. The location of each tadpole was recorded and pooled over five min time intervals across a 30 min evaluation. To ensure that increased occupancy of the target quadrant could not be due to simple inactivity, every 10 minutes, the light pattern rotated 90Â° in a clockwise direction (see diagrams, top of graph). Following initial preference evaluation, tadpoles received red aversion training by punishing individuals in red quadrants with a 1.2 mA electric shock. Conditions were identical to the initial preference phase; lights were rotated 90Â° every 10 minutes during a 30 minute session. Tadpoles were subjected to four identical training sessions followed by 90-minute rest periods under blue light. Following training, tadpoles were reassessed for light preference exactly as in the initial phase, with no quadrants being punished, and showed a significant difference from untrained behavior, spending more time under blue quadrants (2way repeated measure ANOVA P<0.001). Nâ€Š=â€Š24, error bars indicate Â±1 SEM. Red line represents no preference for low intensity red light or high intensity blue light. (B) During each of the 4 training phases, tadpoles strongly avoid the punishing red half of the dish. (C) Examination of the first five minutes of the initial training session reveals that tadpoles move to non-punishing quadrants within the first 30 seconds (6 individual tadpoles shown). (D) Movement rates show an increased â€˜exploratoryâ€™ phase during the first 30 minutes of the trial (initial preference phase) but remain relatively constant across training and post-training sessions. Nâ€Š=â€Š24 for A, B, D and 6 individuals for C, error bars indicate Â±1 SEM. Red line in (A) represents no preference for low intensity red light or high intensity blue light.

References [+] :

Al-Imari, Sight of conspecifics as reward in associative learning in zebrafish (Danio rerio). 2008, Pubmed