Note that this reference documentation is identical to the help that is displayed in MATLAB when you type “help ft_statfun_pooledT”.
FT_STATFUN_POOLEDT computes the pooled t-value over a number of replications. The idea is that you compute a contrast between two conditions per subject The t-values are pooled over subjects and compared against the pooled pseudo-values. Since according to H0 the expected t-value for each subject value is zero, the difference between the pooled t-value and the pseudo-value (which is set to zero) is a fixed-effects statistic. The computation of the difference between pooled t-values can be repeated after randomly permuting the t-values and pseudo-values within the subjects. Each random permutation gives you an estimate of the difference. The random permutations build up a randomization distributin, against which you can compare the observed pooled t-values. The statistical inference based on the comparison of the observed pooled t-values with the randomization distribution is not a fixed-effect statistic, one or a few outlier will cause the randomization distribution to broaden and result in the conclusion of "not significant". Use this function by calling one of the high-level statistics functions as [stat] = ft_timelockstatistics(cfg, timelock1, timelock2, ...) [stat] = ft_freqstatistics(cfg, freq1, freq2, ...) [stat] = ft_sourcestatistics(cfg, source1, source2, ...) with the following configuration option cfg.statistic = 'ft_statfun_pooledT' Configuration options that are relevant for this function are cfg.ivar = number, index into the design matrix with the independent variable See FT_TIMELOCKSTATISTICS, FT_FREQSTATISTICS or FT_SOURCESTATISTICS for details. For low-level use, the external interface of this function has to be [s,cfg] = ft_statfun_pooledT(cfg, dat, design); where dat contains the biological data, Nsamples x Nreplications dat must contain fourier representations. design contains the independent variable (ivar), Nreplications x Nvar