Getting Insight on Animal Behaviour through Interactive Visualization of Multiple T-Maze Ensembles

Information

  • Publication Type: Bachelor Thesis
  • Workgroup(s)/Project(s):
  • Date: August 2020
  • Date (to): 20. August 2020
  • Matrikelnummer: 00827420
  • First Supervisor: Kresimir Matkovic

Abstract

behaviourists and ethologists study cognitive abilities such as learning and memory in rodents to get a better understanding of how similar processes in humans proceed. Often such studies are based on experiments of rodents placed and observed in a Multiple T-Maze. There, the path of the animals are recorded as they move inside the maze, and the resulting trajectories are then analysed. State-of-the-art analysis is based on descriptive parameters and standard statistics where one trajectory at a time is analysed. Usually it is not possible to examine multiple animal paths simultaneously. Together with experts on the field we abstracted the typical work-flow of such analyses and developed an interactive visual analytics tool, with the goal to facilitate the experts’ work and enable a deeper and novel understanding of the learning ability and decision making in rodents. After giving an overview of related works and computer-aided analysis tools in the beginning, the analysis demands and task-breakdown is presented, followed by an Explanation of the data acquisition process, data preprocessing and aggregation. The underlying data structure will be explained as well. The developed analysis tool — the T-Maze Explorer — supports multiple, linked views, which support several traditional methods of visualizations, as well as two newly proposed visualizations fitted to meet the experts’ analysis demands. The first view — the T-Maze View — displays all trajectories of an ensemble with additional options such as highlighting the return path. The purpose of the second view — the Gate-O-Gon view — is to extract information from the trajectories on how often returns in the path occurred and between which parts of the maze these occurred. This information is depicted in a compact and informative novel visualization. The purpose of the T-Maze Explorer is to enable its user to easily find patterns in the data and identify irregular behaviour while inspecting a single path, multiple or the whole trajectory ensemble simultaneously. This thesis provides an insight on how the proposed visualizations were developed, the T-Maze Explorer’s characteristics and benefits as well as it’s limitations. Lastly, a brief excerpt is given on how the T-Maze Explorer could be extended in the future.

Additional Files and Images

Additional images and videos

Additional files

Weblinks

No further information available.

BibTeX

@bachelorsthesis{Bechtold2020,
  title =      "Getting Insight on Animal Behaviour through Interactive
               Visualization of Multiple T-Maze Ensembles",
  author =     "Fabrizia Bechtold",
  year =       "2020",
  abstract =   "behaviourists and ethologists study cognitive abilities such
               as learning and memory in rodents to get a better
               understanding of how similar processes in humans proceed.
               Often such studies are based on experiments of rodents
               placed and observed in a Multiple T-Maze. There, the path of
               the animals are recorded as they move inside the maze, and
               the resulting trajectories are then analysed.
               State-of-the-art analysis is based on descriptive parameters
               and standard statistics where one trajectory at a time is
               analysed. Usually it is not possible to examine multiple
               animal paths simultaneously. Together with experts on the
               field we abstracted the typical work-flow of such analyses
               and developed an interactive visual analytics tool, with the
               goal to facilitate the experts’ work and enable a deeper
               and novel understanding of the learning ability and decision
               making in rodents. After giving an overview of related works
               and computer-aided analysis tools in the beginning, the
               analysis demands and task-breakdown is presented, followed
               by an Explanation of the data acquisition process, data
               preprocessing and aggregation. The underlying data structure
               will be explained as well. The developed analysis tool —
               the T-Maze Explorer — supports multiple, linked views,
               which support several traditional methods of visualizations,
               as well as two newly proposed visualizations fitted to meet
               the experts’ analysis demands. The first view — the
               T-Maze View — displays all trajectories of an ensemble
               with additional options such as highlighting the return
               path. The purpose of the second view — the Gate-O-Gon view
               — is to extract information from the trajectories on how
               often returns in the path occurred and between which parts
               of the maze these occurred. This information is depicted in
               a compact and informative novel visualization. The purpose
               of the T-Maze Explorer is to enable its user to easily find
               patterns in the data and identify irregular behaviour while
               inspecting a single path, multiple or the whole trajectory
               ensemble simultaneously. This thesis provides an insight on
               how the proposed visualizations were developed, the T-Maze
               Explorer’s characteristics and benefits as well as it’s
               limitations. Lastly, a brief excerpt is given on how the
               T-Maze Explorer could be extended in the future.",
  month =      aug,
  address =    "Favoritenstrasse 9-11/E193-02, A-1040 Vienna, Austria",
  school =     "Research Unit of Computer Graphics, Institute of Visual
               Computing and Human-Centered Technology, Faculty of
               Informatics, TU Wien ",
  URL =        "/research/publications/2020/Bechtold2020/",
}