As the doc says,
...All data in the input file must be numeric. dlmread does not operate on files containing nonnumeric data, even if the specified rows and columns for the read contain numeric data only.
As you can tell, csvread has the same limitation since it's a wrapper around dlmread
Also the error shows you it's the first line in the file containing the string 'Devices\n' that it barfed over.
Use textscan or textread or importdata instead...
ADDENDUM
Or, if you're adamant about using csvread, save the data w/o the header information into another file and process that one.
One additional alternative would be to read (I presume there is one) the .xls file directly w/ xlsread which will give the text into as well as the data.
ADDENDUM 2
If all you need is the array of data of the five columns, my preferred approach (being a traditionalist/minimalist in not using cell arrays where aren't needed) would be
d=dir('*.csv');
for i=1:length(d)
m{i}=textread(d(i).name,'','headerlines',5);
end
The above presumes the format of each is identical in having precisely five header lines preceding the data array.
