ppsp-team / HyPyP

The Hyperscanning Python Pipeline
BSD 3-Clause "New" or "Revised" License
75 stars 42 forks source link

PLV in negative value #55

Closed ihgumilar closed 3 years ago

ihgumilar commented 3 years ago

Hi Hypyp team,

As we see the visualization that it shows a blue line, which indicates a negative correlation (red line is for positive correlation). However, for Phase Locking Value (PLV) analysis, I believe that the range of the score is between 0 and 1. If so, how can the visualization give a blue line, which indicates a negative value (which is not in the range of PLV analysis) ? Please advise.

Thanks IG

phoebsc commented 3 years ago

Hi,

Thank you for raising the issue. Would you point out the code/data that has negative PLV? PLV should be positive indeed, as demonstrated by the computation here.

Note that in the "getting_started" notebook, the connectivity values that were visualized were zscored, as in https://github.com/ppsp-team/HyPyP/blob/6445a02bc04ff91d4ebca91c8d8093524b18f868/tutorial/getting_started.py#L175 and line 28 in the notebook

ihgumilar commented 3 years ago

Hi,

Thanks for the reply. I should understand better about the cohen's D indicates there by now.

Also, regarding Comparing Inter-brain connectivity values to random signal in line 44 in notebook, the result says

Using a threshold of 18.512821 stat_fun(H1): min=inf max=inf Running initial clustering Found 1 clusters Permuting 4999 times...

However, when I plotted with the following code, it did not show any line that indicated a significant value of the above cluster. Could you give me some help ?

viz.viz_2D_topomap_inter(epo1, epo2, statscondCluster.F_obs_plot, 
                         threshold='auto', steps=10, lab=True)

image

Thanks

deep-introspection commented 3 years ago

@ihgumilar It should be solved now. Let us know if you get another problem.