SDM calculations

Linear model analyses: meta-regression

This procedure calculates linear meta-regressions, with their statistical significance based on Monte Carlo randomizations. You can specify a filter for subgroup analyses. The following estimates are returned:

- 0: intercept, i.e. prediction from the minimum value of the regressor (which is scaled to 0).

- 1: prediction from the maximum value of the regressor (which is scaled to 1).

- 1m0: slope, i.e. difference between both predictions above.

Notes

Variables are automatically scaled to have values between 0 and 1.

Please be aware that meta-regressions should be generally understood as exploratory and their threshold should be lower than the usual threshold, e.g. 0.0001-0.0002 instead of 0.001.

To calculate a meta-regression

Press the button [Linear model]

or:

Select [Linear model] in the Statistics menu, to open the following dialog:

Command-line and batch usage

lm regressor, filter

Example:

AdultsAgeEffect = lm age, adults

Log

The p = 0.001 thresholds (for positive and negative differences) obtained after each randomization are saved in a text file called (name_of_the_regression).log, e.g. MyRegression_1.log. This is useful for studying the stability of the threshold.

References

Radua J and Mataix-Cols D. Voxel-wise meta-analysis of grey matter changes in obsessive-compulsive disorder. Br J Psychiatry 2009; 195:393-402 .

The authors | IoP King's College London | NeuroImageN