Skip to content

One-sample t-test

In this this tutorial we will perform a one-sample t-test at each voxel in the brain. This will tell us if and where in the brain our group responds to a specific task condition or experimental contrast on average.

Here we will check if there is an overall effect of the task (i.e. contrast con_0009.nii - for a reminder of first-level contrasts, check the previous page).

Specifying the model

  1. Navigate to derivatives/second_level and make a folder for this analysis. Make an empty directory where you will save your results. Name it something meaningful to you, e.g. one_sample_ttest_task.
  2. Select Specify 2nd level from the SPM menu.
  3. In the pop-up batch editor window, select your newly created output directory by clicking Directory and navigating to derivatives/second_level/one_sample_ttest_task in the selection box.
  4. Define your statistical model by selecting Design One-sample t-test (this is selected by default).
  5. Select Scans Specify....
  6. Now, using the selection window recursively filter for contrast con_0009.nii. To do this, navigate to derivatives/first-level via the left-hand side panel. In the filter box, type in ^con_0009.nii and click the Rec button. You should see 40 files selected in the bottom window. Double check that the correct contrast and subjects have been selected. Confirm selection by pressing Done.
  7. Optional: You can also include covariates and/or nuisance variables in your model. These can be specified under Covariates and Multiple covariates. Your covariates should be formatted as columns with rows corresponding to each participant. Always make sure the order in which you input your partipant’s contrasts in the previous step is the same as the order in your covariates list.
  8. From the drop-down menu panel, select SPM Stats Model estimation.
  9. Navigate to Model estimation in the left-hand panel of the batch window.
  10. Press Select SPM.mat Dependency Factorial design specification: SPM.mat file OK.
  11. From the drop-down menu panel, select SPM Stats Contrast manager.
  12. Within the Contrast manager, click on Select SPM.mat Dependency Model estimation: SPM.mat file OK.
  13. You can now start specifying your contrasts of interest in Contrast sessions. We’ll include two t-contrasts, one looking at activations and one looking at deactivations. Select Contrast sessions New: T-contrast.
  14. Give your contrast an informative name, Name Specify.... We’ll name this contrast activations.
  15. Specify your contrast weight, Weights vector Specify... 1.

  16. Now click on Contrast Sessions and add another T-contrast calling it deactivations and giving it a weight of -1.

    F-contrasts

    You can also specify an F-contrast with a contrast weight of 1 to look at both activations and deactivations within one contrast.

  17. Optional: You can also add contrasts exploring the effects of your covariates by specifying contrast weights for the corresponding columns of the design matrix.

  18. When you’re ready, save your batch and press to run your analysis.

Unless you have specified additional covariates, the design matrix for your model should have one column denoting the group effect and look like this:

Viewing the results

To view the results of your analysis, return to the main SPM window and select Results, then select the SPM.mat file corresponding to your analysis. In our case, this will be in derivatives/second_level/one_sample_ttest_task. Choose a contrast to view, e.g. positive.

SPM will now let you select masking and multiple comparisons correction. Select the following in the SPM window:

  • Apply masking none
  • P-value adjustment to control FWE
  • P-value 0.05
  • Extent threshold (voxels) 0
A bit more about SPM results options
  • Apply masking - this option allows you examine your results within a specific region, specified either from a contrast, predefined image, or atlas. To interrogate all voxels in the brain, select none.
  • P-value adjustment to control - this lets you select whether you want to display the results as uncorrected for multiple comparisons(none) or with a FWE voxel/peak-level correction (FWE).
  • P-value - here you can specify your p-threshold.
  • Extent threshold (voxels) - this allows you set a threshold of the minimum number of contiguous voxels in a cluster to be displayed. This can be used in cluster correction.

After going through these steps, SPM will display the results as an activation map and summary table:

The SPM Graphics window shows you the activation map displayed on a glass brain (top left corner) shown in a standardised space with dark blobs representing clusters over the chosen threshold. You can also view your design matrix and the contrast you are currently viewing (top right corner). The bottom part of the results window, show summary statistics table for your results.

The columns in the results table show:

  • cluster-level - the chance (p) of finding a cluster with this many(kE) or a greater number of voxels, corrected (FWE or FDR)/uncorrected for search volume,
  • peak-level - the chance (p) of finding (under the null hypothesis) a peak with this or a greater height (T- or Z-statistic), corrected (FWE or FDR)/uncorrected for search volume,
  • x, y, z (mm) - coordinates in MNI space for each maximum.

You can also display your statistical maps on a different image. To do so, from the drop down menu in the SPM results window which shows overlays, select sections. In the pop-up, navigate to your the directory where SPM is saved and select canonical/avg305T1.nii which is standard template in MNI space. This will display your results as a heatmap on a standard brain template: