Parameters for TEprepare

What values are recommended for the parameter cfgTEP.

• predicttimemin_u/max_u: this is the range of the assumed delay u between any two channels in your data. TRENTOOL will estimate TE for each of these assumed values and check for which u the TE becomes maximal - this u will then be used as the reconstructed interaction delay delta. Note that a wider range will increase the computational time, it is therefore recomended to use (physiologically) plausible values/values that are plausible for your application.
• actthrvalue: This is a threshold value, if the actual ACT is higher for individual trials, these trials will be excluded from the analysis. The value you put here should be based on the filtering of the data prior to TE analysis**: if you highpass filtered your data at 10 Hz and have a sampling rate of 1200Hz, you shouldn't find any autocorrelation above 120 samples. Thus, you may use 120 as a threshold here.
• maxlag: maximum lag in samples used to calculate the ACT for individual trials, have the number of sample points is a good rule of thumb
• minnrtrials: minimum no. trials needed for analysis, if the number of trials falls below this threshold because of trials being rejected due to a high ACT, analysis will abort (should at least be set to 12 to enable premutation statistics over trials)
• ragtaurange: range for the embedding delay tau used by the Ragwitz criterion to determine the optimal embedding - 0.2 to 0.5 are recommended values for standard MEEG data sets (this is in units of the ACT, i.e., round(0.3*ACT) is the number of samples used as tau for a trial)
• ragdim: range for the embedding dimension d used by the Ragwitz criterion to determine the optimal embedding - 2 to 12 are recommended values for standard MEEG data sets
• repPred: This parameter determines how many data points are used for optimization of the embedding parameters by the Ragwitz criterion. TRENTOOL will use the first repPred points in each trial to optimize embedding parameters. This number should be as high as possible (depending on the values you chose for cfgTEP.actthrvalue, fgTEP.ragdim, cfgTEP.ragtaurange, see also TRENTOOL ERROR: Not enough points in timeseries for current analysis settings for a discussion of these parameters).

See also How does the Ragwitz criterion optimize embedding parameters? for further information on the Ragwitz criterion and necessary parameters.

---

Go back to the FAQ page.