T-tests for symptom mapping (fast)

Lesion to symptom mapping performed on a prepared matrix. T-tests are performed using each column of the matrix to split the behavioral scores in two groups. If var.equal=TRUE the Welch test is performed instead. This function relies on TTfast, a compiled version to run on thousands of voxels (about 60 faster then regular t-tests in R).

lsm_ttestFast(lesmat, behavior, var.equal = TRUE,
  alternative = "greater", checkAssumptions = TRUE, showInfo = TRUE,
  ...)

Arguments

lesmat

binary matrix (0/1) of voxels (columns) and subjects (rows).
behavior

vector of behavioral scores.
var.equal

logical (default=TRUE) should the variance between groups considered equal (t-test) or unequal (Welch test).
alternative

(default='greater') Sets the expected relationship between voxel value and behavior. By default voxels with zero are not lesioned, and behavior is expected to be higher, thus alternative='greater'. If the relationship in your data is inverted, use alternative='less', and if you don't have a relationship hypothesis data, use alternative='two.sided'.
checkAssumptions

(default=TRUE) Check how many voxels violate the t-test assumptions (heteroscadsticity and/or normality).
showInfo

logical (default=TRUE), display time-stamped info messages
...

other arguments received from lesymap.

Value

List of objects returned:

  • statistic - vector of statistical values

  • pvalue - vector of pvalues

  • zscore - vector of zscores

Examples

{
set.seed(123)
lesmat = matrix(rbinom(200,1,0.5), ncol=2)
set.seed(123)
behavior = rnorm(100)
result = lsm_ttestFast(lesmat, behavior)
}
#>  
#> 07:56:57     checking variance homogeneity... 0 voxels failed (0%)
#> 07:56:57     checking distribution normality... 0 voxels failed (0%)