correlations.to.adjacencies {brainwaver} | R Documentation |
Given a correlations thingy as produced by const.cor.list
,
produce a list of adjacency matrices fiddled to have a preferred number of edges
Actually this is not quite possible, but come as close as choose.thresh.nbedges
will allow us. A functional parameter allows us to say things like produce the graphs with n log n edges where n is the number of nodes
correlations.to.adjacencies(correlations, edge.func) ideal.wavelet.levels(brain) distance(x,y,z)
correlations |
a list of correlation matrices produced by const.cor.list |
edge.func |
a function to mention the way to choose the number of edges given the number of nodes in the graph. In the companion scripts files, the small-limit is used and by default edge.func=(function(x){x*log(x)}) |
brain |
matrix containing the data time series. Each column of the matrix represents one time series. |
x |
x coordinate |
y |
y coordinate |
z |
z coordinate |
Functions produced to manipulate better nice outputs of the package
correlations.to.adjacencies |
Description of 'comp1' |
ideal.wavelets.levels |
number indicating up to each wavelet scale it is possible to go given the length of the time series |
disctance |
the euclidean distance in 3D |
John Aspden, external collaborator of the brainwaver package
S. Achard, R. Salvador, B. Whitcher, J. Suckling, Ed Bullmore (2006) A Resilient, Low-Frequency, Small-World Human Brain Functional Network with Highly Connected Association Cortical Hubs. Journal of Neuroscience, Vol. 26, N. 1, pages 63-72.
data(brain) brain<-as.matrix(brain) # WARNING : To process only the first five regions brain<-brain[,1:5] n.levels<-4 wave.cor.list<-const.cor.list(brain,n.levels=n.levels) adj.mat<-correlations.to.adjacencies(wave.cor.list,edge.func=(function(x){x*log(x)}))