NITRC BetA-Series COrrelation Forum: open-discussion

The second temporal derivative of HRF in estimating the "amplitude of effects"

Gao F — Wed, 28 Apr 2021 4:49:02 GMT

In the paper 'BASCO: a toolbox for task-related functional connectivity', estimation of the "amplitude of effects" is proposed as:
[color=#ff0000]β = sign(β1) * sqrt(β1^2 + β2^2 + β3^2), [/color][color=#ff0000]where, the first beta-value relates to the canonical HRF, [/color][color=#ff0000]the second and third beta-values relate to the first and second [/color][color=#ff0000]temporal derivative, respectively.[/color]

[color=#000000]I was wondering how to set or [size= 10px]create[/size] the [color=#ff0000]β3[/color]. Then I check the BASCO.m, where in the 1270 to 1272[/color] lines of the script, it is commented:
[color=#000000]% Estimating the "amplitude" of the effects at each voxel = sign(V1).*sqrt(V1.^2+V2.^2)[/color]
[color=#ff0000]% where V1 is the canonical effect contrast volume, and V2 is the temporal derivative[/color]
[color=#ff0000]% effect contrast volume. [Calhoun (2004)][/color]

[color=#ff0000][color=#ff0000]β3[/color] seems not included in the [/color][color=#000000]formula in the comment[/color][color=#000000], and in the following line 1282 of the script:[/color]
datmat = sign(datmatA).*sqrt(datmatA.^2+datmatB.^2+[color=#ff0000]datmatC[/color].^2)
[color=#ff0000][color=#ff0000]
[/color][/color]
The '[color=#ff0000]datmatC[/color]' seems corresponding to 'β3', but it more likely a disperion derivatives while not the second temporal derivative.

My question is, in which way the BASCO estimating the "amplitude of effects", and by which way the second temporal derivative of HRF could be used, or if I have just misunderstood the script?
[color=#ff0000][color=#ff0000]
[/color][/color]

Long processing time at "Model specification and estimation"

Luiz Kobuti Ferreira — Sun, 13 Sep 2020 11:08:00 GMT

Hi,

I have a question about processing time during the step "Model specification and estimation"

I am testing BASCO to analyze task-related functional connectivity. During the step "Model specification and estimation" I noticed that the time BASCO needs to process each participant increases as it processes data from each participant. For instance, the first participants were processed in less than 10min while when getting to the 10th participant, it took almost one hour per participant and then, at participant number 17, it took more than 3 hours per participant!

Is this dramatic increase in processing time per participant when looping through participants during "Model specification and estimation" a normal behaviour?

Thanks in advance,
Luiz Kobuti Ferreira

Data preprocessed in FMRIPrep?

jensadler — Mon, 19 Nov 2018 18:56:27 GMT

Hello!

Thank you for this toolbox! It is incredibly helpful!

I have a question about data that can be run through the toolbox. One of our studies has been preprocessed through fMRIPrep (https://fmriprep.readthedocs.io/en/stable/index.html). Can we use the preprocessed output in BASCO. or does the data need to be preprocessed in SPM?

Thanks!
Jenny

Error after running ana.m

abdulqawi alarefi — Thu, 22 Jun 2017 7:42:12 GMT

Dear Martin Göttlich,

Thanks for this awesome toolbox. After executing most of the steps, I got an error after trying to run ana.m script.

Loading file F:\ana\fmri_softwares\others\BASCO_v2\BASCO\tutorial\ana.m.
Reference to a cleared variable AnaDef.
Error in BASCO>pushbuttonmodelspecest_Callback (line 200)
NumSubj = AnaDef.NumSubjects;
Error in gui_mainfcn (line 95)
feval(varargin{:});
Error in BASCO (line 20)
gui_mainfcn(gui_State, varargin{:});
Error in
matlab.graphics.internal.figfile.FigFile/read>@(hObject,eventdata)BASCO('pushbuttonmodelspecest_Callback',hObject,eventdata,guidata(hObject))
Error while evaluating UIControl Callback

Sincerely,

Abdulqawi

Usability GUI and estimated model

Johann Kruschwitz — Wed, 08 Feb 2017 15:54:17 GMT

Hey,
very nice toolbox congratulations! I just played around a little and noticed that it might be a useful addon allwing to select the folders of already estimated models via the GUI (i.e., I did all the model specification and estimation for a very large sample outside of BASCO and would only like to perfom everything starting from ROI-based model estimation (estimate model on mean ROI time courses)). As far as I picked it pup from the BASCO.m file all model estimations are pretty much SPM defaults - or am I mistaken? I changed the code but it is quite messy with all the variable assignments from the ana.def reference :)

Do you have a new version with these specifications on hold?

All the Best,
Johann

Empty conditions running GLM models

L Tang — Tue, 31 Jan 2017 4:35:42 GMT

Dear BASCO author and users,

I am trying to use BASCO to quickly generate the beta series for my analysis. For most participantsm there are 4 conditions in my analysis for most participants. But for some, one or two conditions would be empty for certain runs. That conditions are not of interest but has to be modeled. In this case, BASCO would error out. I am wondering if there is an easy way to ask BASCO to flexibly ignore the empty condition or I can only specify less conditions for these.

Thanks.
Lingfei

RE: Error

Lisanne Jenkins — Mon, 28 Nov 2016 18:00:31 GMT

Hi,
I made the changes to the file basco_checkestimation on line 16 and to line 225 in BASCO.m but I still receive the following error message for some subjects:

Retrieving data from ROI 57 using summary function mean ...
Summarizing data with summary function: mean
Re-estimating covariance
Re-estimating whitening filter
Initialising parameters : ...done
Covariance estimate : ...voxelwise
Model : estimation
Temporal non-sphericity (over voxels) : ...estimation
ReML Block - 1
ReML Iteration : 1 ...3.410752e+25
ReML Iteration : 2 ...2.828835e+03
ReML Iteration : 3 ...4.421080e+01
ReML Iteration : 4 ...1.613961e+06
ReML Iteration : 5 ...4.180238e+01
ReML Iteration : 6 ...4.352357e+06
ReML Iteration : 7 ...4.636968e+01
ReML Iteration : 8 ...4.689308e+01
ReML Iteration : 9 ...1.634696e+04
ReML Iteration : 10 ...5.011224e+01
ReML Iteration : 11 ...3.862712e+03
ReML Iteration : 12 ...8.503537e+01
ReML Iteration : 13 ...5.833874e+02
ReML Iteration : 14 ...8.889501e+02
ReML Iteration : 15 ...1.863922e+02
ReML Iteration : 16 ...4.274574e+02
ReML Iteration : 17 ...1.672121e+03
ReML Iteration : 18 ...5.110463e+05
ReML Iteration : 19 ...4.828098e+01
ReML Iteration : 20 ...4.704126e+01
ReML Iteration : 21 ...6.273130e+05
ReML Iteration : 22 ...4.188783e+01
ReML Iteration : 23 ...7.261460e+03
ReML Iteration : 24 ...5.100982e+01
ReML Iteration : 25 ...8.972958e+03
ReML Iteration : 26 ...1.505456e+06
ReML Iteration : 27 ...1.283097e+02
ReML Iteration : 28 ...2.008960e+03
ReML Iteration : 29 ...2.246124e+03
ReML Iteration : 30 ...3.719424e+03
ReML Iteration : 31 ...7.797202e+06
ReML Iteration : 32 ...6.124269e-03
ReML Block - 2
ReML Iteration : 1 ...6.028477e+24
ReML Iteration : 2 ...6.179355e+02
ReML Iteration : 3 ...9.867692e+02
ReML Iteration : 4 ...2.847287e+02
ReML Iteration : 5 ...3.228575e+02
ReML Iteration : 6 ...2.212133e+02
ReML Iteration : 7 ...3.051855e+04
ReML Iteration : 8 ...1.003979e+01
ReML Iteration : 9 ...1.422214e+03
ReML Iteration : 10 ...5.317306e+01
ReML Iteration : 11 ...6.749793e+00
ReML Iteration : 12 ...1.624683e+06
ReML Iteration : 13 ...1.148070e+05
ReML Iteration : 14 ...1.237820e+11
ReML Iteration : 15 ...3.307808e+10
ReML Iteration : 16 ...2.169239e+01
ReML Iteration : 17 ...2.275552e-01
ReML Iteration : 18 ...2.411023e-02
ReML Iteration : 19 ...8.187992e-03
ReML Block - 3
ReML Iteration : 1 ...8.628086e+24
ReML Iteration : 2 ...1.992343e-01
ReML Iteration : 3 ...1.852284e-01
ReML Iteration : 4 ...2.820160e-01
ReML Iteration : 5 ...1.189589e-01
ReML Iteration : 6 ...3.375698e-01
ReML Iteration : 7 ...9.604632e-01
ReML Iteration : 8 ...3.161417e-01
ReML Iteration : 9 ...4.307677e-01
ReML Iteration : 10 ...2.068754e-01
ReML Iteration : 11 ...3.447115e-01
ReML Iteration : 12 ...7.117137e-01
ReML Iteration : 13 ...2.499299e-01
ReML Iteration : 14 ...2.532627e-02
ReML Iteration : 15 ...1.525531e-01
ReML Iteration : 16 ...5.933029e-01
ReML Iteration : 17 ...6.832739e-01
ReML Iteration : 18 ...2.374132e+00
ReML Iteration : 19 ...3.896834e-01
ReML Iteration : 20 ...1.498764e-02
ReML Iteration : 21 ...5.230396e-01
ReML Iteration : 22 ...6.657243e-01
ReML Iteration : 23 ...1.587401e+00
ReML Iteration : 24 ...2.196471e-02
ReML Iteration : 25 ...3.874127e-01
ReML Iteration : 26 ...9.265903e-02
ReML Iteration : 27 ...6.606886e-01
ReML Iteration : 28 ...1.192551e+00
ReML Iteration : 29 ...9.113033e-01
ReML Iteration : 30 ...5.524773e-02
ReML Iteration : 31 ...7.925981e-01
ReML Iteration : 32 ...6.872270e-01
ReML Iteration : 33 ...5.076939e-01
ReML Iteration : 34 ...1.386557e-01
ReML Iteration : 35 ...5.097154e-01
ReML Iteration : 36 ...2.239037e-02
ReML Iteration : 37 ...8.979627e-02
ReML Iteration : 38 ...7.340414e-02
ReML Iteration : 39 ...1.083122e-01
ReML Iteration : 40 ...2.579397e-01
ReML Iteration : 41 ...2.568578e-01
ReML Iteration : 42 ...1.010973e+00
ReML Iteration : 43 ...3.614601e+00
ReML Iteration : 44 ...1.057727e+00
ReML Iteration : 45 ...1.504028e-01
ReML Iteration : 46 ...2.351308e-02
ReML Iteration : 47 ...1.543476e-01
ReML Iteration : 48 ...3.993500e-01
ReML Iteration : 49 ...2.469874e-01
ReML Iteration : 50 ...4.375068e-01
ReML Iteration : 51 ...2.231699e-01
ReML Iteration : 52 ...4.365207e-01
ReML Iteration : 53 ...7.082969e-01
ReML Iteration : 54 ...4.793554e-01
ReML Iteration : 55 ...9.630564e-02
ReML Iteration : 56 ...3.928655e-01
ReML Iteration : 57 ...1.110299e+00
ReML Iteration : 58 ...1.810996e+00
ReML Iteration : 59 ...1.407834e+00
ReML Iteration : 60 ...1.211868e-01
ReML Iteration : 61 ...3.618090e-01
ReML Iteration : 62 ...2.167922e-01
ReML Iteration : 63 ...1.829130e-01
ReML Iteration : 64 ...8.644111e-01
ReML Block - 4
ReML Iteration : 1 ...6.866919e+24
ReML Iteration : 2 ...5.011138e+00
ReML Iteration : 3 ...2.151095e-01
ReML Iteration : 4 ...5.406613e-02
ReML Iteration : 5 ...8.375658e-01
ReML Iteration : 6 ...1.334521e+00
ReML Iteration : 7 ...7.013734e-02
ReML Iteration : 8 ...1.285255e+00
ReML Iteration : 9 ...1.387347e+00
ReML Iteration : 10 ...1.316738e+00
ReML Iteration : 11 ...4.913568e-02
ReML Iteration : 12 ...4.446727e+00
ReML Iteration : 13 ...3.673783e-01
ReML Iteration : 14 ...5.815593e-01
ReML Iteration : 15 ...6.456046e-01
ReML Iteration : 16 ...8.945228e-01
ReML Iteration : 17 ...1.561681e+00
ReML Iteration : 18 ...4.379552e-01
ReML Iteration : 19 ...4.910517e-03
ReML Block - 5
ReML Iteration : 1 ...2.162717e+24
ReML Iteration : 2 ...8.876681e+03
ReML Iteration : 3 ...2.196735e+01
ReML Iteration : 4 ...3.817622e+01
ReML Iteration : 5 ...1.270137e+05
ReML Iteration : 6 ...4.325122e+01
ReML Iteration : 7 ...5.862207e+04
ReML Iteration : 8 ...3.096524e+02
ReML Iteration : 9 ...1.539215e+03
ReML Iteration : 10 ...7.389998e+05
ReML Iteration : 11 ...5.132941e+02
ReML Iteration : 12 ...2.444608e+02
ReML Iteration : 13 ...1.067493e+04
ReML Iteration : 14 ...2.140501e+06
ReML Iteration : 15 ...4.237760e-03
Initialising parameters : ...computingIndex exceeds matrix dimensions.

Error in spdiags (line 114)
a((len(k)+1):len(k+1),:) = [i i+d(k) B(i+(m>=n)*d(k),k)];
Error in pr_estimate (line 128)
s = spdiags(1./sqrt(diag(s)),0,nScan,nScan);
Error in pr_estimate (line 345)
SPM = pr_estimate(SPM,marsY);
Error in mardo_5/estimate (line 72)
SPM = pr_estimate(SPM, marsY);
Error in BASCO>pushbuttonmodelspecest_Callback (line 423)
E = estimate(D,Y); % estimate model based on ROI
summary
Error in gui_mainfcn (line 96)
feval(varargin{:});
Error in BASCO (line 20)
gui_mainfcn(gui_State, varargin{:});
Error in
@(hObject,eventdata)BASCO('pushbuttonmodelspecest_Callback',hObject,eventdata,guidata(hObject))
Error while evaluating uicontrol Callback

Thank you in advance for your help!

RE: Troubles displaying correlation matrix

Martin Göttlich — Mon, 29 Aug 2016 9:47:41 GMT

Dear Isabelle,

you may send me the BASCO output-file so that I' m able to reproduce the problem: martin.goettlich@neuro.uni-luebeck.de.

Best,

Martin

Troubles displaying correlation matrix

Isabelle Ripp — Wed, 24 Aug 2016 13:37:46 GMT

Hey,

first of all; great toolbox!! After having done the tutorial, I adjusted the anadef.m to my data and all worked fine. I calculated the ß-Series and wanted to compute the correlation matrix (after I computed the connectivity matrix by selecting "Product-moment correlation"), but matlab kept on showing me the same error (see below). I am new to matlab, so it might be something obvious, but nevertheless, I would be very happy, if one could help me fixing the error.

Thanks ahead

Isabelle

Calculated correlation coefficients from beta-series.
[color=#ff0000]Undefined function 'imshow' for input arguments of type 'double'.[/color]

[color=#ff0000]Error in InspectCorrMatrix>DisplayMatrix (line 91)[/color]
[color=#ff0000]imshow(NWM,'DisplayRange',[],'Parent',handles.axes1);[/color]
[color=#ff0000]Error in InspectCorrMatrix>InspectCorrMatrix_OpeningFcn (line 30)[/color]
[color=#ff0000]DisplayMatrix(handles);[/color]
[color=#ff0000]Error in gui_mainfcn (line 220)[/color]
[color=#ff0000]feval(gui_State.gui_OpeningFcn, gui_hFigure, [], guidata(gui_hFigure), varargin{:});[/color]
[color=#ff0000]Error in InspectCorrMatrix (line 19)[/color]
[color=#ff0000]gui_mainfcn(gui_State, varargin{:});[/color]
[color=#ff0000]Error in BASCO>pushbuttoncorrelationmatrix_Callback (line 598)[/color]
[color=#ff0000]InspectCorrMatrix(handles.anaobj);[/color]
[color=#ff0000]Error in gui_mainfcn (line 95)[/color]
[color=#ff0000]feval(varargin{:});[/color]
[color=#ff0000]Error in BASCO (line 20)[/color]
[color=#ff0000]gui_mainfcn(gui_State, varargin{:});[/color]
[color=#ff0000]Error in @(hObject,eventdata)BASCO('pushbuttoncorrelationmatrix_Callback',hObject,eventdata,guidata(hObject))[/color]
[color=#ff0000]Error while evaluating uicontrol Callback [/color]

RE: Error

Jerry Zhu — Thu, 11 Aug 2016 17:43:47 GMT

[color=#000000]Thanks a lot for the code snippet, Ralf. You are awesome![/color]

also the initial source of info
http://marsbar.sourceforge.net/faq.html#i-get-errors-using-the-spm-reml-estimation-for-fmri-designs-can-i-try-something-else

There is also a function called autocorr from econ toolbox. Type "which autocorr" to see which autocorr is working.
if necessary, one can copy @mardo_2/autocorr.m

[i]Originally posted by Ralf Veit:[/i][quote]Hi,

this error was reported several times in the marsbar forum. Matthew Brett suggested to use K. Worsleys fmristat instead of the autocorrelation implementation in SPM. For that you must insert the following code in the file basco_checkestimation on line 16.

D = autocorr(D, 'fmristat', 2);

I have also noticed that sometimes I get the error "Check number of ROIs in txt-file"
I changed line 225 in BASCO.m from
if length(thenames)~=ROINum
to
if size(thenames,1)~=ROINum

and it worked.

Best
Ralf[/quote]