10.07.2015 Views

Deconvolution Analysis of FMRI Time Series Data - Waisman ...

Deconvolution Analysis of FMRI Time Series Data - Waisman ...

Deconvolution Analysis of FMRI Time Series Data - Waisman ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

(Can you prove this claim?)The command -tts myFit causes the tted time series to be written to le myFit.1D.You can plot the input time series and the tted time series in separate graph windowswith the following commands (suppressing the rst 6 points):1dplot -ignore 6 029 029 008.1D1dplot -ignore 6 myFit.1DHowever, if you wish to plot the tted time series on top <strong>of</strong> the <strong>FMRI</strong> time series data,the following commands can be used to combine the two time series into a single le (e.g.,Overlay.1D) for plotting:Script for Plotting Fitted <strong>Data</strong> on top <strong>of</strong> Actual <strong>Data</strong>1dtranspose 029 029 008.1D 029 029 008.1Dt1dtranspose myFit.1D myFit.1Dtcat 029 029 008.1Dt myFit.1Dt > Overlay.1Dtrm -f Overlay.1D1dtranspose Overlay.1Dt Overlay.1Drm -f 029 029 008.1Dtrm -f myFit.1Dtrm -f Overlay.1Dt1dplot -ignore 6 -one Overlay.1DYou should see the tted time series (in red) plotted on top <strong>of</strong> the original time series (inblack). This result is similar to that which you will obtain in Example 2.3.1 (q.v.) usingthe <strong>Deconvolution</strong> plugin.Example 1.4.3.2 Multiple Linear Regression using Estimated Motion ParametersIn this Example, we repeat the above analysis, but will include in the full model theestimated motion parameters. The reason for doing so is to account for variation in themeasured response due to subject motion. This should help both to reduce false positives(by subtracting out the eect <strong>of</strong> stimulus correlated motion) and to increase true positives(by reducing the measurement noise estimate).The rst step is to estimate the subject motion. This can be done using program3dvolreg, e.g.:3dvolreg -base 34 -1Dfile v2.motion.1D v2:time+origThis will write to le v2.motion.1D the estimated motion parameters: roll, pitch, yaw,dS. dL, and dP, as 6 columns <strong>of</strong> numbers. Now, these separate columns can be entered asindividual stimulus functions, as indicated below:37

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!