3dMVM bogus results error

AFNI Message Board
1

Hi AFNI Gurus,

After running a 3dMVM analysis, the output ended with the following:


Congratulations! You have got an output ANCOVA-Faces-DES-only+tlrc.

Warning message:
Numerical variables NOT centered on 0 (i.e., likely bogus results): Age, WK2_DESB_RS

However, I set up my script to mean center the quantitative variables - that is the values provided in the “-qVarCenters” option are the averages of the Age and W2_DESB_RS variables, respectively:


	3dMVM -prefix ANCOVA-DES-only -jobs 4                                				\
		-bsVars  "Site*Sex+Age+WK2_DESB_RS"                 	\
		-wsVars "Face"                                                  					\
		-qVars "Age,WK2_DESB_RS"                           	\
		-qVarCenters '34.66,1.36'          		                        		\
		-SS_type 3                                                       					\
		-GES																				\
		-resid Residuals																	\
		-num_glt 12                                                       					\
		-gltLabel 1 'Fear' -gltCode 1 'Face : 1*Fear'    									\
		-gltLabel 2 'Neut' -gltCode 2 'Face : 1*Neutral'    								\
		-gltLabel 3 'Fear-Neut' -gltCode 3 'Face : 1*Fear -1*Neutral'    					\
		-gltLabel 4 'Fear-DES' -gltCode 4 'Face : 1*Fear WK2_DESB_RS : '   		 			\
		-gltLabel 5 'Fear-DES-H' -gltCode 5 'Face : 1*Fear WK2_DESB_RS : 5.24'   		 	\
		-gltLabel 6 'Fear-DES-L' -gltCode 6 'Face : 1*Fear WK2_DESB_RS : 0.00'   		 	\
		-gltLabel 7 'Neut-DES' -gltCode 7 'Face : 1*Neutral WK2_DESB_RS : '   		 		\
		-gltLabel 8 'Neut-DES-H' -gltCode 8 'Face : 1*Neutral WK2_DESB_RS : 5.24'   		\
		-gltLabel 9 'Neut-DES-L' -gltCode 9 'Face : 1*Neutral WK2_DESB_RS : 0.00'   		\
		-gltLabel 10 'F-N-DES' -gltCode 10 'Face : 1*Fear -1*Neutral WK2_DESB_RS :'   		\
		-gltLabel 11 'F-N-DES-H' -gltCode 11 'Face : 1*Fear -1*Neutral WK2_DESB_RS : 5.24'	\
		-gltLabel 12 'F-N-DES-L' -gltCode 12 'Face : 1*Fear -1*Neutral WK2_DESB_RS : 0.00'	\
		-dataTable @Faces_Table_MVM-DES-only.txt

Given this, what could cause this warning and how should it be addressed?

Thanks for your time and assistance.
Best,
Lauren

2

Lauren,

That’s a benign warning, and you can safely ignore it. It comes from the R package “afex” that warns the user of the situation when the overall mean is not removed from a quantitative variable. In your case it was your intention to center around a value other than the overall mean. So everything is fine.

3

Great! Thanks, Gang!