I need some help with arrayfun

I have this data from a mice experiment:

• Reaction Time
• Response Made (which can be 0,1 or 2)
• Visual contrast (one of 9 possible values)

I would like to use arrayfun to calculate the mean reaction time for each of the 9 values of visual contrast when the response made was 2.

Each set of data is written in a separate Matlab table.

Thank you.

[jasmine]

This is how my data looks like.

First column - Reaction Time

Second column - Response Made

Third column - Visual contrast

[the cyclist]

Can you upload the data itself, or a small sample, instead of an image?

[jasmine]

Attached is the data

[jasmine]

https://docs.google.com/spreadsheets/d/17z7j9qBwWHu1j_8xHIyoQ3Tb5qk0rZaLnWO4nG1rE5Q/edit?usp=sharing

This contains the data as well

[the cyclist]

Is the data already in MATLAB? Can you upload a MAT file?

The reason I suggest that is that (a) there are many ways to load data into MATLAB, which can give subtly different MATLAB variables, and (b) then we can test code directly on your data, to see if our algorithm works for you.

[jasmine]

it is in MATLAB, but it is structured in a lot more complicated form and I think it won't be very helpful to update it like that

[the cyclist]

Well, the main information I am trying to get at is the data types and how you are storing.

You make the statement "Each set of data is written in a separate Matlab table". That suggests to me that you have three variables of data type table. But your screen shot shows three columns of a single variable of data type double.

I'm just trying to sort this out, so that I don't suggest a solution that doesn't actually work for how your variables are stored.

[jasmine]

Sorry, yeah, they are normally stored as single double type.

Answer 1

[Steven Lord]

I don't think arrayfun is the tool you want to use here. Instead groupsummary seems like a better fit.

[jasmine]

I used arrayfun to calculate the probability to obtain a specific response (0, 1 or 2) for each of the 9 contrast values. But now that I had to deal with the reaction time as well, it got too complicated. Do you have any idea how I could use groupsummary for this?

[jasmine]

Pressuming that the response made is always = 2 (also making it a bit easier this way), how can I get the mean reaction time for each of the 9 visual contrasts?

[Steven Lord]

Take a look at the "Multiple Grouping Variables" and "Multiple Grouping Vectors for Vector Input" examples on the documentation page to which I linked. You should be able to adapt these to your data.

Instead of computing the mean Weight for people grouped by Gender and whether or not they are Smokers, you could compute the mean ReactionTime for items grouped by Response and Contrast.

Answer 2

[the cyclist]

I believe that groupsummary is for data stored in tables only. You can do what you want with

[sortedResponses,~,idx] = unique(R(:,2));
meanTimes = accumarray(idx,R(:,1),[],@mean);

Here, R is your original data, and meanTimes will have the mean values for the sortedResponses.

You could also just do

mean(R(R(:,2)==2),1)