BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Cambridge Analysts' Knowledge Exchange
SUMMARY:Estimating whole brain dynamics using spectral clu
stering - Yi Yu (University of Cambridge)
DTSTART;TZID=Europe/London:20151104T160000
DTEND;TZID=Europe/London:20151104T170000
UID:TALK61850AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/61850
DESCRIPTION:Spectral clustering is a computationally feasible
and model-free method widely used in the identific
ation of communities in networks. In this work\, w
e introduce a data-driven method\, namely Network
Change Points Detection (NCPD)\, which detects cha
nge points in the network structure of a multivari
ate time series\, with each component of the time
series represented by a node in the network. NCPD
consists of three parts: spectral clustering allow
s us to consider high dimensional time series wher
e the dimension of the time series is greater than
the number of time points (N > T)\; the principal
angles allows for estimation of the change in ter
ms of network/graph structures across time without
prior knowledge of the number or location of the
change points\; permutation and bootstrapping meth
ods are used to perform inference on the change po
ints. NCPD is applied to various simulated data se
ts as well as to a resting state functional Magnet
ic Resonance Imaging (fMRI) data set. The results
illustrate the ability of NCPD to observe how the
network structure changes over the time course. T
he new methodology also allows us to identify comm
on functional states across subjects. Finally\, t
he method promises to offer a deep insight into th
e large-scale characterisations and dynamics of th
e brain.
LOCATION:MR14\, Centre for Mathematical Sciences
CONTACT:Adam Kashlak
END:VEVENT
END:VCALENDAR