Project Details
Projekt Print View

Parallax in Electric Field Imagery

Subject Area Electronic Semiconductors, Components and Circuits, Integrated Systems, Sensor Technology, Theoretical Electrical Engineering
Measurement Systems
Term from 2019 to 2021
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 427854804
 
Final Report Year 2024

Final Report Abstract

Within the 18-month funding period of the project we addressed if and how weakly electric fish influence the shape of the self-generated electric field with a focus on the question if self-motion can be interpreted in terms of an active recruitment of depth information through electrical parallax. On the biological side, we conditioned weakly electric fish in a novel 2-AFC paradigm to electrically discriminate objects based on size and distance. Test trials demonstrated that decisions were predominantly made based on the relative size differences. However, stereotyped probing motor acts (va-et-vient movements) were not found to be linked to the performance. As part of the project, we started to analyze the temporal capacity of the working memory required to solve the task. While data is still being obtained, the results confirm work on zebrafish, documenting a span of 1-2 seconds. Despite the unexpected absence of va-et-vient motor patterns, we have been able to further our conceptual understanding of the potential role of such stereotyped motor-patterns beyond the use for depth perception. This led to a re-focus with respect to experimental conditions where we now investigate the potential role of motor patterns in the formation of a spatial representations for navigation. This is made possible through the implementation of markerless tracking as well as the use of neural-network based EOD discrimination. In a continuation of the funded work on the results from a different lineage of weakly electric fish, we here explore if va-et-vient motor-patterns contribute to spatial learning as we found that these motor patterns do occur both in task acquisition and in test trials when fish need to discriminate dense and sparse electrical scenes. On the technological side, we addressed if the concept of visual parallax can be transferred to sensor systems in fluids based on electric fields, such that a formal description of “electric parallax” is obtained. We set up an analytical description of an abstraction of the electrical field of the fish which was used for a simulation study that resulted in a definition of “electric parallax” based on peak traces formed by changing voltage profiles on a moving sensor line. This lead to a novel definition of an “electric aperture angle” and to a further extension which allows definition of “electric parallax” also for the non-moving sensor line. These results are currently tested in hardware. To link technological and biological work, we are currently addressing how active field shaping can aid in extracting sensory information. In this approach, we also investigate how the temporal sequence of sensor images may support information quality. This will be coupled to the ongoing investigation of the SO movements in electrosensory learning, where the technological knowledge will aid in the quantification of the sensory flow encountered.

Publications

 
 

Additional Information

Textvergrößerung und Kontrastanpassung