UCI Dataset problem with Neural network nprtool

I'm trying to resolve this classification problem from UCI Learning:

https://archive.ics.uci.edu/ml/datasets/Daily+and+Sports+Activities

The thing is that it has a lot of files...each one is an instance of the network so I need put them all together as an input.

How can I pass all those files to the network?

Also:

http://kilyos.ee.bilkent.edu.tr/~billur/publ_list/hbu10.pdf

Here is the procedure that the developers did for their network for every text file.(Get some features of the signals that are 45, and then apply PCA)

I know how to get the feautures in only one file, how can I do this to all of them.

And also, PCA is available in MATLAB?

Thanks and cheers!

From the paper:

"Each of the five sensor units has nine sensors; thus, 45 signals are available for each 5-sec time window. We calculate the following 26 features for each signal: the minimum and maximum values, the mean value, variance, skewness, kurtosis, 10 equally spaced samples from the auto correlation sequence, first five peaks of the discrete Fourier transform of the signal and the corresponding frequencies. As a result, 1,170 (= 45×26) features are available for each 5-sec window for each activity. All features are normalized to the interval [0,1] to be used for classification"