Table 1.

Princeton's MVPA toolbox classifier performance (in percent) distinguishing pitch from timbre conditions using four different masks

Subject	ROI mask thresholded to 2000 voxels	ROI mask	Step size correlated voxel mask intersected with 2000 voxel mask	Step size correlated voxel mask	Mean classifier performance (%) for each subject	SD
1	73	70	71	68	70.5	2.08
2	45	43	47	38	43.25	3.86
3	49	53	50	47	49.75	2.50
4	55	52	51	57	53.75	2.75
5	65	63	66	64	64.5	1.29
6	69	72	69	63	68.25	3.77
7	56	56	61	60	58.25	2.63
8	86	80	70	74	77.5	7.00
9	56	56	55	56	55.75	0.50
10	62	66	60	63	62.75	2.50
Mean classifier performance (%) for each mask	61.6	61.1	60	59	60.4	1.16
SD	12.17	11.11	8.91	10.34	10.3

Values in bold indicate best classifier performance for each subject.