.MAT to .Dat issue

I am trying to convert the data from a series of 4 .mat MEG trails (4 sensors x651 recording) to a .dat file for analysis. When I try to use dlmwrite not all sensors are included in the .dat file and when i try to convert using -ascii the .dat file alters the format of the data from a single ERP to continous data spikes. Does anyone know of another way to create my .dat file. Kind regards

[dpb]

Why can't you just load the .mat file and analyze the date?

What's the need for another file containing the same thing?

Show the actual code and some of the data and what you got versus what you think you should have gotten...

[Robert Mayger]

the program (spm) that I am using requires the waveform information to be seperated into a .dat file inorder to be read, I will attach the entire .mat file, besa_sourcewaveforms_all.waveforms is the variable that I am trying to extract, I have also added figures for how the .dat files appear in my program, they are labelled accordingly to the process I used.

I dont have any extensive code but the code I did use followed the formats previously suggested on mathworks

dlmwrite code - dlmwrite('SUB01.dat',[waveforms])

ascii code - save('SUB01.dat', 'waveforms', '-ascii')

[dpb]

>> whos -file SUB01_besa_source.mat
  Name                          Size             Bytes  Class     Attributes

  besa_sourcewaveforms          1x1              29448  struct              
  besa_sourcewaveforms_all      1x15            430064  struct              

>> load SUB01_besa_source.mat
>> besa_sourcewaveforms
besa_sourcewaveforms = 
  struct with fields:

        structtype: 'besa_sourcewaveforms'
            labels: {'SD-1'  'SD-2'  'SD-3'  'SD-4'}
         waveforms: [651×4 double]
    numberoftrials: 141
         condition: 'ISI200_R_M'
      samplingrate: 1000
              type: 'GRA'
         latencies: [1×651 double]
          datafile: 'C:\Documents and Settings\All Users\Documents\ADAPT\SUB06 (28-07-2010) New analysis\Sub006_MH_GRANDAVERAGE_av_new.fsg'
         modelname: 'C:\Documents and Settings\All Users\Documents\ADAPT\SUB06 (28-07-2010) New analysis\ECD_Solution\Sub006_MH_4ECD_Solution_new.bsa'
         headmodel: 'MEG spherical'
    regularization: 0.0100
           filters: '1 - 100 Hz  NF:50'
>> 

It's a struct with a field waveforms, the other is an array of those.

dlmwrite('yourfilename',besa_sourcewaveforms.waveforms)

will write the array to the file as .csv; I forget the default precision (6 significant digits, maybe???)

BTW, dlmwrite has been deprecated; TMW suggests using writematrix instead.

>> dlmwrite('besa_sourcewaveforms.csv',besa_sourcewaveforms.waveforms(1:10,:))
>> type besa_sourcewaveforms.csv

0,0,0,0
0,0,0,0
0.017103,0.012915,0.015769,0.0070932
0.073439,0.050682,0.041433,0.014394
0.13735,0.084326,0.019121,-0.00096016
0.14264,0.066736,-0.050773,-0.0085693
0.072304,-0.024836,-0.064302,0.060108
-0.010435,-0.18761,0.079103,0.2102
0.006966,-0.39497,0.3814,0.36595
0.20388,-0.58668,0.74595,0.45235
>>
>> besa_sourcewaveforms.waveforms(1:10,:)
ans =
         0         0         0         0
         0         0         0         0
    0.0171    0.0129    0.0158    0.0071
    0.0734    0.0507    0.0414    0.0144
    0.1373    0.0843    0.0191   -0.0010
    0.1426    0.0667   -0.0508   -0.0086
    0.0723   -0.0248   -0.0643    0.0601
   -0.0104   -0.1876    0.0791    0.2102
    0.0070   -0.3950    0.3814    0.3660
    0.2039   -0.5867    0.7460    0.4523
>> 

Illustrates does work as advertised...so, we can't see what you actually did, but code snippet posted doesn't have meaning on its own as there is no variable waveforms so it's unknown how you created it. Dereference the struct and all should be well...