About the project
To understand the brain at the cellular level, we need large-scale neural interfacing. Developing more advanced prosthetic devices will facilitate the next generation of brain machine interfaces (BMIs).
This project will focus on high bandwidth implantable brain processing techniques using Utah Arrays with over 1000 recording channels.
Observing a larger number of recording sites generates significant data for interpretation. Increased data throughput has led to challenges with computational efficiency, particularly for on-chip algorithms in which power budgets cannot be readily increased.
The key challenge with high-bandwidth intracortical processing is associated with the overall computation cost. This is also translated into the power consumption and the area of the implant. Therefore, this demands a paradigm shift from conventional brain data processing approaches.
As a researcher on this project, you'll have the opportunity to:
- design low-cost and highly reliable deep learning architectures suitable for real time and on-chip spike sorting
- develop an intelligent on-chip processor integrable into the sensing module for real time behavioural decoding of different regions of the brain
- validate the algorithms (simulation and fabricated) using the recorded and resynthesized in-vivo test data to demonstrate the brain-implantation potential for the BMIs
You'll also spend time gaining many scientific skills working in the our modern laboratory facilities. You'll develop your knowledge and collaborate with a strong multidisciplinary team of academics in the Digital Health and Biomedical Engineering group and the Analogue and Biomedical Electronics Group at UCL, with Professor Demosthenous.