User's Guide‎ > ‎Analysis‎ > ‎

6 - Statistics

What is it good for? 

These calculations provide first hints as to GFP and/or topographic differences between experimental conditions. Ttests can also be applied to IS or on the output of the fitting procedure. The example bellow is on ERPs.

 

-          Tools >>> Statistics

-          Choose Set of Files

-          Add new group of files >>> load the singl_avr.eph of one condition from the folder with the grand_avr (remember: it’s only them which are eventually baseline-corrected between subjects!)

-          Add new group of files comprising the singl-avr.eph of a second condition

-          Output file: *.ep is ok

-          Choose output folder….

-          Click on “2 parameters”

o        In “Variables to Test” you can choose your electrodes or the GFP or whatever (with certain limitations e.g. TANOVA only takes electrodes)

o        Choose your test (e.g. paired t-test for the electrodes and/or GFP)

o        Threshold the p-value eventually to 5% and 11 TF (with the latter corresponding to ~20ms of “effect” duration)

o        The various *.ep output files you can easily read into your favourite statistics and graphics program (time scales in ms can be easily created from the TF values by dividing each TF by your EEG sampling frequency and multiplying by 1000 afterwards….)

 

 

Tip: The output of the TANOVA gives you global Dissimilarity values (= 1 minus p-value). However, the observed effects can either stem from topographic differences between conditions or latency shifts of responses between conditions! As TANOVA is an index of topographic modulations, its result should be consistent with the segmentation/fitting results; you can thus use TANOVA results to drive the segmentation/fitting procedure.






Comments