It seems they are taking the natural log of the mean expression and calculating the difference.

mean.fxn <- function(x) {
    return(log(x = mean(x = expm1(x = x)) + 1)) 
}

data.1 <- apply(X = data.use[features, cell.use1, drop = FALSE],MARGIN = 1,FUN = mean.fxn)
data.2 <- apply(X = data.use[features, cell.use2, drop = FALSE],MARGIN = 1,FUN = mean.fxn)

FC <- (data.1 - data.2)

You can have a look at the code here.