TDT: the figure of group confusion matrix in the publication

1

Hi everyone,

I am curious to know how to plot a group confusion matrix like in the TDT publication Figure 5A . The figure is very useful for decoding visualization, especially for multiclass classification. However, I can’t find such related descriptions in these demo scripts.
I have tried ROI-based decoding analysis with my own dataset for 4-conditions, but got only one value showing the accuracy to classify the 4-conditions. The problem is how can I get the exact value of each cell in the confusion matrix.
Should I change the default model parameters (train: ‘-s 0 -t 4 -c 1 -b 0 -q’; test: ‘-q’)? any suggestions?

Thank you in advance.
Luffy

2

Hey Luffy,

just use
cfg.results.output = {‘confusion_matrix’}
if you only want the confusion matrix, or e.g.
cfg.results.output = {‘accuracy_minus_chance’, ‘confusion_matrix’}
if you like both output measures.

For more, see “help transres_confusion_matrix”.

Happy decoding!
Kai

3

Hi Kai,
Thank you so much for your prompt reply.
The simple solution really works now.

Best wishes,
Luffy