Sample fMRI Block Design Analysis using FSL
This tutorial describes how to analyze a simple fMRI dataset. To complete this tutorial you will need:
- FSL installed on your computer (included with Lin4Neuro).
- Download and extract the sample dataset that includes the fMRI data (in NIfTI format), an anatomical scan (also T1) and text files that describe the timing.
Consider a simple experiment where we ask an individual to press their left hand at some timepoints, their right hand at other times, and to simply rest at other moments in time. In this situation one would expect that the left motor cortex and right cerebellum will be activated following right hand movement. This tutorial shows how to analyze this dataset. Below we provide a sample dataset from the GSU/GT Center for Advanced Brain Imaging.
The image below illustrates our study. The vertical axis shows time, in 1.92 second fMRI volumes (the picture shows the first 6 minutes). The red and green bars along the top of the image show when the individual presses their left (red) or right hand (green) - note that they press the same hand continuously for 12 seconds (a 'block' design). The red and green lines show the predicted blood flow signal for regions that are activated following left and right hand movements. Note that we expect the images to become brighter several seconds after an individual starts a task (the blood flow effect is sluggish).
Analyzing the dataset
- From the command line, type "cd ~/tutorial" to change to the folder with the tutorial data. Next start FSL by typing 'fsl &'.
- By typing fsl, a user interface (the one with a fossil fish on its top, the logo for 'FSL') appears. This interface contains several options, for this tutorial we will use BET and FEAT.
- Start FEAT: From the main FSL menu, press "FEAT FMRI analysis" This will open the FEAT interface. (The following steps are described for use in FEAT version 5.98)The FEAT interface has two top buttons. The following analysis is a first level analysis (within one single subject).The FEAT interface five tabs across the top portion of the screen, oriented in a horizontal line. The fmri analysis can be set by defining the parameters within each one of those buttons, in a left to right order: Data, Pre-Stats, Stats, Post stats, and Registration.
- Data Tab: Number of inputs is set to one. Press select 4D data. Browse to the folder in which the 4D data file is stored and select the corresponding file (e.g. fmriblocks009) After selecting the 4D file (make sure to press the "OK" button in the file dialog), the number of volumes will automatically be displayed (in our case 302). Next, enter the TR: in our case 1.92. Set the delete volumes to 0. Set the high pass filter cutoff to 48s (we have 12-second long blocks, so twice the on-off period is 48s). You can set the output directory: you must be able to write to this directory. If you leave it blank, it will write the results in the same folder where the 4D data is. FEAT will write the results in a folder called "4Ddataname.feat". In our case fmriblocks009.feat. If you re-run the analysis, it will create a second folder called fmriblocks009.feat+, and so on.
- Pre-stats Tab: Slice timing correction is probably not useful (we have a fast TR and a block design). Keep the default Motion correction (FMRIB's Linear Image Registration Tool - MCFLIRT). Keep "Brain Extraction" checked. Set "Spatial smoothing FWHM" set to 8mm (between x2-x3 the raw resolution of the functional imaging). Make sure to keep "Intensity Normalization" (equivalent to Grand Mean Scaling) unchecked. Keep "Temporal filtering Highpass" checked.
- Stats Tab: The onsets timings of our design is intricate so press the "Full Model Setup" button. - Clicking full model setup opens a "General Linear Model" interface. Within this interface, follow the steps: - The number of original EV (original EVS - explanatory variables, or simply put: conditions) is set to 2, because we are interested in the response to the "left" and "right" conditions. - Each EV is setup separately. - The basic shape of the wave form that describes the stimulus that we wish to model is defined by a custom file expliciting the time in seconds of the onset of the stimulus, how much time (also in seconds) the stimulus last, and the value of the input. This requires a text file containing a matrix under which the columns are the variables stated above (i.e., first one time of onset, second one the duration, and the thrid one the intensity) and the rows represent the stimuli.
Post_stats Tab: Keep all the defaults.
Registration Tab: To keep things simple, we will only use the fMRI data for normalization (see the event-related design tutorial to see how to use the T1 scan to aid normalization), so keep the the "Main structural image" box uncheck. Make sure that the "standard space" normalizations use 12 DOF (12 parameters: rotation, zoom, shear, translation in 3 dimensions) - this is often not set by default, but works much better.
Press the 'Go' button on the bottom left. A FEAT report window appears - at first it will report "Started at ...STILL RUNNING", but slowly it will populate teh subpages until it says "Finished at...".
A web page will be generated displaying the results.
- Setting Up Conditions: This is set as follows: set the "Number of original EVs" to 2. This will create two sub-Tabs in the EVs tab ('1','2': one for each condition). For the '1' tab, set the EV name to "L" and choose the "Custom (3 column format)" option from the "Basic Shape" pull-down menu, finally set the filename (press the folder icon to browse for the filename) to LBlock.tab", make sure "Add temporal derivative" is unchecked (we have a block design). For the '2' tab, set the EV name to "R" and choose the "Custom (3 column format)" option from the "Basic Shape" pull-down menu, finally set the filename (press the folder icon to browse for the filename) to RBlock.tab", make sure "Add temporal derivative" is unchecked (we have a block design).
- Setting Up contrasts: Once these parameters have been entered click on 'Contrasts & F-tests' tab. In theory, we could create quite a few statistical comparisons, for example, here are five possible contrasts of interest:
- OC1: Find what is active for left motor presses: title= left>, EV1= 1, EV2= 0
- OC2: Find regions more active for left than right motor: title= left>right, EV1= 1, EV2= -1
- OC3: Find what is active for right motor presses: title= right>, EV1= 0, EV2= 1
- OC4: Find regions more active for right than left motor: title= left>right, EV1= -1, EV2= 1
- OC5: Find regions more active for any motor activity title= motor>rest, EV1= 1, EV2= 1
- Alternatively, we could only look at our two favorite conditions (left > right and right > left).
- Press Done to continue.