brainGraph_GLM specifies and fits a linear model at each vertex for a
given vertex measure (e.g. degree) or at the graph-level (e.g.,
global efficiency). Given a contrast matrix, it will calculate the
associated statistics.
The summary method prints the results, only for which
\(p < \alpha\); you may change this to the FDR-adjusted or permutation
p-values via the function argument p.sig.
The plot method plots the GLM diagnostics (similar to that of
plot.lm). There are a total of 6 possible plots,
specified by the which argument; the behavior is the same as in
plot.lm. Please see the help for that function.
brainGraph_GLM(g.list, covars, measure, con.mat, con.type = c("t", "f"),
X = NULL, con.name = NULL, alternative = c("two.sided", "less",
"greater"), alpha = 0.05, level = c("vertex", "graph"),
permute = FALSE, N = 5000, perms = NULL, long = FALSE, ...)# S3 method for bg_GLM
summary(object, p.sig = c("p", "p.fdr", "p.perm"),
contrast = NULL, digits = max(3L, getOption("digits") - 2L),
print.head = TRUE, ...)
# S3 method for bg_GLM
plot(x, region = NULL, which = c(1L:3L, 5L), ...)
A list of igraph graph objects for all subjects
A data.table of covariates
Character string of the graph measure of interest
Numeric matrix specifying the contrast(s) of interest; if only one contrast is desired, you can supply a vector
Character string; either 't' or 'f' (for t or
F-statistics). Default: 't'
Numeric matrix, if you wish to supply your own design matrix
(default: NULL)
Character vector of the contrast name(s); if con.mat
has row names, those will be used for reporting results (default:
NULL)
Character string, whether to do a two- or one-sided test
(default: 'two.sided')
Numeric; the significance level (default: 0.05)
Character string; either vertex (default) or
graph
Logical indicating whether or not to permute group labels
(default: FALSE)
Integer; number of permutations to create (default: 5e3)
Matrix of permutations, if you would like to provide your own
(default: NULL)
Logical indicating whether or not to return all permutation
results (default: FALSE)
Other arguments passed to brainGraph_GLM_design
A bg_GLM object
Character string specifying which p-value to use for displaying
significant results (default: p)
Integer specifying the contrast to plot/summarize; defaults to showing results for all contrasts
Integer specifying the number of digits to display for p-values
Logical indicating whether or not to print only the first
and last 5 rows of the statistics tables (default: TRUE)
A bg_GLM object
Character string specifying which region's results to
plot; only relevant if level='vertex' (default: NULL)
Integer vector indicating which of the 6 plots to print to the
plot device (default: c(1:3, 5))
An object of class bg_GLM containing some input-specific
variables, in addition to:
A numeric matrix; a copy of the design matrix
A numeric vector or matrix of the outcome variable
A data table with an entry for each vertex (region) containing statistics of interest
A character vector of Study.ID's removed due to incomplete data (if any)
A list containing: null.dist (the null distribution of maximum statistics), thresh (the statistic value corresponding to the \(100 \times (1 - \alpha)\)th% percentile of the null distribution)
The plot method returns a list of ggplot objects
The input list of graphs g.list must not be nested; i.e., if you have
multiple groups, they will have to be combined into one list. See the code in
the Examples below.
A data.table of covariates is required input; the first column
must be named Study.ID. Additionally, all graphs must
have a name attribute (at the graph level) which matches the
Study.ID for a given subject. If you create the design matrix
X yourself, you still must supply the covariates table so that
subjects can be correctly matched with their network data.
Both t- and F-contrasts are allowed. You may supply a matrix to the
argument con.mat. If you supply a multi-row matrix and you choose
con.type="t", then statistics will be calculated for each contrast
individually. If you choose con.type="f", in the result data table,
ESS stands for "extra sum of squares", the additional variance
explained for by the model parameters of interest (as determined by the
contrast matrix). Finally, the standard error in these tables is the sum of
squared errors of the full model.
Finally, you can calculate permutations of the data to build a null distribution of the maximum statistic, to provide control over false positives. The permutation strategy is that of Freedman & Lane (1983), and is the same as that in FSL's randomise.
Freedman D & Lane D (1983). A nonstochastic interpretation of reported significance levels. J Bus Econ Stat, 1(4):292-298.
Nichols TE & Holmes AP (2001). Nonparametric permutation tests for functional neuroimaging: A primer with examples. Human Brain Mapping, 15(1):1-25.
Winkler AM, Ridgway GR, Webster MA, Smith SM, Nichols TE (2014). Permutation inference for the general linear model. NeuroImage, 92:381-397.
Other GLM functions: GLMdesign,
GLMfit, mtpc
Other Group analysis functions: Bootstrapping,
IndividualContributions,
MediationAnalysis, NBS,
brainGraph_permute, mtpc
# NOT RUN {
conmat <- matrix(c(0, 0, 0, 1), nrow=1)
rownames(conmat) <- 'Control > Patient'
## Note that I concatenate the graphs from each group's 6th threshold
g.lm <- brainGraph_GLM(g.list=do.call(Map, c(c, g))[[6]],
covars=covars.all[tract == 1],
measure='strength', con.mat=conmat, alt='greater',
permute=TRUE, long=TRUE)
# }
# NOT RUN {
## Save objects and then to multipage PDF
lmPlots <- plot(x)
ggsave('lmPlots.pdf', lmPlots)
## Save all the GLM sub-objects from MTPC analysis
res.mtpc <- mtpc(...)
glmPlots <- lapply(res.mtpc$res.glm, plot, which=1:6)
ml <- marrangeGrob(glmPlots, nrow=1, ncol=1)
ggsave('glmPlots.pdf', ml, width=8.5, height=11)
# }
Run the code above in your browser using DataLab