Vision System Enables First Automated Learning and Testing Chamber with Real-Time Feedback
A new methodin examining test subject behavior has been developed by the Tufts Centerfor Regenerative and Developmental Biology partnering with Wireless Techniques. Given that manualmethods performed by human researchers can be inaccurate, expensive, and time-consuming,the new automated learning and testing chamber can instead analyze the behaviorof small animals on a 24/7 basis, with several experiments running at the same time.As a result, greater insight will be achieved in the area of learning and memory.
Using human researchers to observe test subject behaviorduring research experiments can be both time consuming and expensive. Theavailable manpower for real-time observation is limited, and human observationsare inherently subjective. To address some of these limitations, the Tufts Centerfor Regenerative and Developmental Biology has partnered with WirelessTechniques to develop the first automated learning and testing chamber foranalyzing behavior in small animals. The chamber uses a Cognex In-Sight Microvision system instead of a human researcher to observe the behavior of testsubjects.
Tufts researchers are using the new testing chamber to studythe molecular mechanisms underlying the ability of living things to learn fromtheir environment. Light stimuli are used to train worms and tadpoles onspecific tasks, and the animals are then tested for recall in a variety ofmolecular-genetic and pharmacological experiments. The new tracking systemprovides quantitative data on the subjects’ behavior and performance inlearning tests. As a unique system, it is the first to not only allow trackingof animal movement, but to also provide parallel, independent feedback to eachsubject so they can learn specific tasks. Simple animals such as flatwormsshare many of the same behavioral pathways and neurotransmitters with humanbeings. Accordingly, these animals are often studied to better understand theproperties of memory storage and transmission in tissue. The new chamber makesit possible to test new drug compounds to determine if they impact cognitiveability.
According to Professor Michael Levin, Director of The TuftsCenter for Regenerative and Developmental Biology, “Modern cognitive science isstriving to understand the connection between molecular genetics and theinformation processing mechanisms that give rise to behavior and thought. Thebiomedical aspect of this goal includes the search for drugs that will aidlearning and memory and the understanding of various influences on cognition.”
In a typical experiment, worms will be trained to stay in oravoid specific parts of the dish, or to move at specific rates. Worms that successfullyperform the task will be rewarded by lowered light levels, as worms naturallyprefer the dark.
Until now, studies have been performed manually. However,the manual approach of assessing behavior puts significant limits onexperimental progress. Only a limited number of animals can be analyzed by handdue to manpower and cost limitations. Manual handling may also allow theresults to be affected by the judgment and errors of the person running theexperiment. For example, the lack of consensus on the learning abilities offlatworms has been attributed to the small sample sizes that have been requiredby manual training. Manual methods make it difficult or impossible for otherlabs to replicate results and for other scientists to view the originalexperiment and potentially uncover trends that might have been missed by theexperimenter.
The Tufts Center selected WirelessTechniques to design and build an automated learning and testing chamber thatcould provide real-time feedback without a human researcher. WirelessTechniques (now through its successor-in-interest Boston Engineering Corporation—a product and systems development services firm with its mainoffice in Waltham, MA—which acquired a substantial amount of WirelessTechniques’ assets) designs and builds custom electronic devices andinstrumentation for applications including wireless and wired communications,sensing, and signal processing. Cognex was chosen as the vision system supplierbecause its sophisticated image processing tools could determine the positionof the worms despite complicated shadowing effects created by the movement ofwater in the test chamber.
How the Chamber Works
The chamber consists of 12 cells arranged in a grid, each holding a disposablePetri dish where the worm lives. The environment in each cell is individuallycontrolled by the software depending on the behavior of the animal within. Thelid contains a series of light emitting diodes (LEDs) controlled by a computerthat are used to train worms. A set of four bright LEDs can be set toilluminate a single quadrant of the dish, and barriers prevent the light fromspreading to adjacent quadrants. Red LEDs that cannot be seen by the worms areapplied at all times during experiments so that the vision system can track themotion of the worms without having any effect on the worm’s behavior.Electrodes in the dish allow the experimenter to also provide weak electricalsignals to the animals.
 |
| Topview of the group of four “Experiment Environment Modules” with two of the IlluminationHeads open to show access to the Shock Electrode Holder and Petri Dish. |
Each experiment is controlled by an algorithm written byLevin’s team. First, the position of the worms in their dishes is recorded bythe vision camera, followed by a certain action, such as turning on a light inone quadrant of the dish with the goal of teaching the animal to swim to thelighted quadrant. Next, the position of the worm is once more recorded by the visioncamera. Based on the position (and second-order quantities like speed,direction of movement, etc.) of the worm, another action might be taken such asrewarding the worm by turning down the lights because it swam to the correct quadrant,or turning on a bright light because it did not perform the task properly.These series of measurements and actions can continue until the program reachesa predefined condition (a level of performance indicating that the animal hasunderstood the task to be learned).
Since the system is automated, 12 experiments can be runsimultaneously seven days a week, 24 hours a day without human intervention. Asa result, much larger sample sizes can be achieved, and experiments can also berun for much longer periods. Millions of observation and training cycles can beperformed, creating a level of training far beyond what can realistically beaccomplished by manual methods. The system also provides complete consistencyamong experiments, allowing labs to replicate experiments performed elsewhere,and reduce the amount of noise in the data. Additionally, the vision systemrecords the worms’ motion, meaning it can be easily reviewed and analyzed byother experts over the Internet.
Overcoming the Vision Challenge
“Machine vision was one of the greatest challenges in this automated learningsystem,” said Chris Granata, former President of Wireless Techniques, nowProgram Manager, Wireless and Sensing Technologies at Boston Engineering. “Thewater touching the sides of the dish creates a meniscus that rises and falls.This creates shadows that change over time and are difficult to distinguishfrom the worms. This application requires a vision system with powerful visiontools that are capable of identifying the location of the worm and iscompletely self-contained in a compact package so we can easily increase thenumber of cells. The Cognex In-Sight Micro 1400 was ideal for this applicationbecause of its broad toolset and the fact that the entire system is containedin a 30 mm x 30 mm by 60 mm enclosure.”
 |
Three experiment cells as viewed bythe Cognex Insight Micro-1400. |
In order to reliably differentiate worms from randomlychanging water shadows, images of empty quadrants are captured every 20seconds. The action is accomplished by tracking the worm’s position andcapturing quadrants while they are not occupied by the worm. When the systemcaptures an image of the worm in a quadrant, it subtracts the most recent imageof the same quadrant when it was not occupied by the worm in order to removethe shadows and more accurately determine the position of the worm.
A histogram tool is used to identify and group the lightestcolored pixels, which determine possible positions of the worm. Severalconvolution and morphological filters are used to enhance the image. Forexample, morphological dilation filtering is used to connect white pixels inclose proximity to each other and smooth out edges of white islands. Next, ablob detection tool picks out the three largest groups of light colored pixelsand sorts them in order of size. In almost every case, the largest object isthe worm; however, multiple objects are tracked to address the rare possibilitythat one or more shadows may be larger than the worm.
Conclusion
Scientists hope to use this research to make discoveries about the molecularbasis of memory and to develop the latest nootropic drugs.
“We are using quantitative automated behavior analysistechniques to ask how and where information is encoded and how it can beimprinted upon the regenerating brain by other tissues,” said Levin. “Geneticchanges can be made to the worms and then their learning performance can bemeasured in the chamber in order to understand which genes affect learning andmemory. This chamber also provides a very powerful tool for investigating themechanisms of memory and behavior and for drug screening of new nootropiccompounds designed to treat conditions such as attention-deficit hyperactivitydisorder (ADHD), drug addiction, etc. as well as counteract effects ofneurotoxins and improve cognitive performance. Automating the training andtesting process will enable us to make faster progress by running many moreexperiments on a 24/7 basis instead of just when human experimenters areavailable; moreover, the quantitative data (impossible to obtain with humanobservers) will reveal unprecedented insights into the processes of learningand memory.”
Mark W. Smithers is VP/COO at BostonEngineering Corporation. He is responsible for overseeing general engineering operations support such as facilities, communications, productivity tools, development standards and CAD/CAE systems. Smithers can be reached at 781-314-0714 or info@boston-engineering.com.
| ADVERTISING » | « ADVERTISING |
