CMU9 fMRI brain dataset

The dataset is used mainly in the paper:

"Predicting Human Brain Activity Associated with the Meanings of Nouns" by Mitchell et al.

The details of the dataset can be found in the url below:

The dataset is made available in MATLAB's .mat file whose details are given below:

There are 9 subjects

Totally 360 observations

12 categories, each having 5 exemplars, each having 6 observations. 360 = 12*5*6.

Untitled spreadsheet

When we do animal vs tool, we will have

animals: 'animal' + 'insect' --> 10 exemplar --> 60 observations

tools: 'tool' + 'furniture' --> 10 exemplar --> 60 observations

6 runs (epoch)

Do 6-fold cv --> test:validation:train = 1:2:3

train: 30 observations for each class --> 60 obs in total

validation: 20 obs for each class --> 40 obs in total

test: 10 obs for each class --> 20 obs in total

Next, some codes to extract the information from the data are shown:

we need to make fold_id, x, y, y_label

Untitled spreadsheet

1--"animal"

2--"insect"

3--"tool"

4--"vegetable"

result multiclass classification