Topological Methods in Brain Network Analysis (17w5043)

Arriving in Banff, Alberta Sunday, May 7 and departing Friday May 12, 2017


(University of Western Ontario)

(University of South Florida)

(University of Iowa)


Algebraic topological methods have been used in network analysis and topological data analysis, while spatial graph theory and knot theory have been used mostly in biomolecular interactions. Recent development in persistent topology approach to data analysis gives a rather coarse view on data that often needs much finer analysis on its spatial connectedness. The aim of the proposed workshop is to investigate applicability of topological methods in studying spatial interconnectedness on neuroscience data and other biomolecular data by developing new mathematical techniques for spatial molecular arrangements through already developed topological methods used in biomolecular interactions. In particular we aim to:

(a) Train topologists in understanding neuroimaging

Tutorials on the basics of neuroscience experimental observations and neuroimaging will be conducted by experts in these fields, who are also familiar with the language and methods of topology. These tutorials will be hands-on so that workshop attendees leave fully able to access, and understand, the data produced in these disciplines (tutorial speakers who confirmed their interest in the workshop: Randy McIntosh)

(b) Expose neuroscientists to topological methods used for data analysis

Methods such as persistent topology and mathematical models for spatial graphs and networks have significant potential in studying complex data arising form neuroscience labs. We propose to have general tutorials for these mathematical methods that will be accessible to neuroscientists and other mathematicians (tutorial speakers who confirmed their interest in the workshop: Yuliy Baryshnikov on networks, Carina Curto on topological data analysis in neuroscience, Mariel Vazquez on knot theory).

(c) Software tutorials

Participants will be encouraged to install relevant software before the start of the meeting. However, we will also have times at the beginning of the conference to assist with installation. Some of the software will be installed on the conference site and the conference will provide basic training. Software manuals are often not directed towards a general audience. Since our participants will come from a wide variety of backgrounds, their thoughts on what issues they encountered could be very helpful. We will include advice on how to handle such issues in a website we will create as described below in (e). Software tutorial speakers who confirmed their interest in the workshop: Vidit Nanda and Mikael Vejdemo Johansson.

(d) Create plans for outreach to biologists and neuroimagers

We anticipate that about one third of the workshop participants would be junior scientist including graduate students and postdocs. One of the goals of the workshop is to facilitate interdisciplinary interactions in particular to including creating e possible venues for lasting interdisciplinary communications such as including possibilities for neuroscientists to hosting math graduate students/post-docs in their labs as well as mathematicians hosting neuroscience students/post-docs in their labs. for short of up to a semester long visit.

(e) Creation of accessible tools and training materials

We will develop a webpage to introduce both those with a mathematical background as well as those in the medical field to topological methods in spatial interactions, network and data analysis. This website will include links to: — the recordings by BIRS of the tutorials given in this workshop as well as speaker slides. — participants’ recommended articles, books, websites including online courses related to introductory data analysis, topological methods in network analysis and spatial arrangements, and as well as neuroscience. — available developed software; a significant component of the workshop will be dedicated to learning how to use already available software and to identify the missing components for successful spatial network study. — the human connectome project and other data repositories as suggested by the participants. — links to other related sites such as Institute for Mathematics and its Applications (IMA) special year in Research Network in Applied Algebraic Topology

The success of bioinformatics has been hugely aided by a culture of producing user- friendly, openly accessible, software implementations as a key deliverable in fundamental bioinformatics research. The cleverest algorithms which exist only as theorems and proofs in a paper will not be used by working biologists while simple algorithms with friendly web interfaces have become standard tools for students of molecular biology. We would like to build a consensus around the importance of adopting this user-friendly practices and discuss opportunities to seek funding for a centralized repository of software (in a manner similar to, e.g.: NCBI, EMBnet, NITRC). A natural extension of the goal of providing accessible tools is the provision of accessible, and expertise-appropriate, training.