Decoding geometrical structures from neural population recordings using Machine Learning and Deep Learning

Neurons situated in a brain region called the hippocampus have been discovered to represent information related to numerous aspects of space and navigation. The activity of these neurons is referred to as the cognitive map, a complex brain representation of space, navigation and other associated physical clues. The aim of this project is to design, develop, implement and test artificial intelligence algorithms to decode neural data recorded from the hippocampal formation. The neural data considered is, among others, recorded from rodents freely moving and navigating in a complex environment. The developed algorithms will aim to analyse and better understand how space is
represented in rodents’ and humans’ brains and how the cognitive space representation reacts and adapts to different topologies and different external stimuli. Finally, understanding how cognitive maps dynamics change and, in particular, how spatial memory is negatively affected by time and external stimuli shall lead to a better understanding of neurological dementias such as Alzheimer’s disease.

Faculty Supervisor:

Manu Sasidharan Madhav

Student:

Partner:

École polytechnique fédérale de Lausanne

Discipline:

Life Sciences

Sector:

Education

University:

The University of British Columbia

Program: