Patient N and % Bars for Loc and Lat

[neurokleos]

The localisation vertical bar should show on one side the number of total localising data points (patient N) used to plot the heatmap and on the other side the normalised %

The lateralisation horizontal bar below the axial and coronal brain, perhaps should be divided in 2 halves R and L, each one showing the lateralising data points (patient N) used to plot the heatmap as well as the normalised %

[fepegar]

Assuming this table, how would you calculate those numbers?

	Structure X	Structure Y	Structure Z
Semiology A	0	1	1000
Semiology B	0	2	3
Semiology C	0	50	5
Semiology D	0	12	13
Semiology E	0	0	0

[neurokleos]

With a combined semiology can be tricky, since the Total absolute sample is different for each semio, but talking about one semio at a time this should be doable. The question is: to get to the colour bar showing data from 0 to 100 intensity, we should process the absolute raw numbers first, so we would just need to show up the raw number next to the %, that's it. Similar issue here (not with colourbar though) https://community.tableau.com/thread/254499

[fepegar]

I didn't know you meant just one semiology. That will be a particular case, though.

If we have just one semiology, I think the best would be just showing the raw numbers, so no percentage would be needed anyway.

What about that horizontal bar? By the way, showing a horizontal bar there might be trickier as this is not implemented in Slicer.

[fepegar]

Reading the email you sent to John, I think I understand a bit better what you want for the horizontal bar, but I'm not sure a bar is needed to show two numbers (if that's indeed what you're thinking of).

[neurokleos]

When a neurologist will interrogate the semio-vis tool to have an idea of where might be the seizure onset zone from which her/his patient's epilepsy signs and symptoms originate, he/she will be proceeding with displaying first the onset (ideally) or as early as possible semiology (that is likely to be a single one, e.g. automatism, dialeptic, aura, tonic), this to have a putative map of the seizure onset zone. So the main thing in terms of colourbars is providing the clinician with an idea of the sample size of patients whose localising/lateralising data they are looking at, for each semiology - only this number can give them the sense of the mega-review data and of how much informative and reliable is the putative map.

[fepegar]

I understand, thanks for the explanation.

What I mean is that a colourbar is used to map colours to numbers. If neurologists need numbers and not colours, there's no need for a bar.

[neurokleos]

They need both. The colour is a way to visualise number. An example it comes to my mind is like when you hear a music without knowing its musical notes. For someone who wants to understand the music better it would be nice to know the notes as well.