Puzzle-solving Bacterial Pet: Imaging Platform for Microfluidics-based Reinforced Learning with Motile Bacterial Cells
Summary
This proposal is for developing a programmable staging mount, and an imaging platform for a microfluidics-based reinforced learning hub for motile bacterial cells. By developing a maze traversal challenge, we aim to create different scenarios for chemotactic bacterial colonies to employ their decision-making machinery and navigate their way out of the maze. By isolating successful colonies and progressing them further to solve more complex mazes, we aim to identify strains of bacteria that are good at solving maze traversing puzzles. The potential learning from this could lead to an understanding of cognition, memory and learning in bacterial colonies. In addition, we aim to understand a bacterial colony’s ability to learn from its experience of solving maze traversal puzzles by measuring complexity level of each of the maze designs.
The Team
Dr. Emre Ozer, Principal Research Engineer, Research Division at ARM Ltd.
Emre will contribute towards designing the microfluidics maze and HW/SW system architecture.
Dr. Varun Kothamachu, Postdoc, The Babraham Institute (Signalling ISP)
Varun will co-ordinate the project and contribute towards electronic microcontrollers needed. All members will be involved in overall experiment design, image analysis algorithm and analysing data generated
Dr. Tanya Hutter, Research Fellow, Department of Chemistry at University of Cambridge
Tanya will contribute towards building the microfluidics platform for the conditioning hub.
Dr. Pahini Pandya, Research Associate, University of Cambridge
Pahini will contribute towards the experimental wet lab work needed for growing the bacteria.
Project Outputs
Project Report
Project report and documentation on Github
Project Proposal
Original proposal and application