The brain as a temporal network

My perspective is to view the brain as a system interacting over time. This involves quantifying the brain as a network that changes through time.

Theory

Coming soon

Results

When applying the different methods we have developed we are able to find different single trial behavioural properties correlating with temporal network properties (e.g. fluctuating temporal communities, changes in segregation/integration). We have also identified temporal patterns on resting state data that include a “bursty” temporal pattern of network interactions

Methods

Several different time-varying connectivity methods have been proposed. The two most promising methods proposed are: the jackknife correlation and weighted pearson correlatino by timepoint similarity.

We have also have proposed some metrics to quantify temporal networks and a method for temporal community detection from time series.

Software

Teneto is a python package for temporal network analyses.

TVC benchmarker is a set of simulations that would test different methods ability to track a fluctuating covariance and gives the opportunity for other researchers to submit their method to be tested alongside..

Meta-science

I am also involved and have been involved in a number of more meta-science project where the aim is to quantify how science is done and/or improve scientific practice.

One project demonstrated that the readability of scientific texts is decreasing through time.

Another project used Neurosynth to see how much the different network terminology in neuroimaging literature overlaps