TABLE 3.

ROC numbers.

Cut-off Value	Sensitivity (%)	Specificity (%)
Validated group
−161	100	0
−108	100	6
−99	100	13
−52	100	19
−45	100	25
0	100	31
10	100	38
15	100	44
19	100	50
27	100	56
36	100	63
59	100	69
81	100	75
92	89	75
101	78	75
107	67	75
108	56	75
132	44	75
195	33	75
215	33	81
220	22	81
251	22	88
275	11	88
326	0	88
432	0	94
Quality group
−108	100	5
−99	100	9
−52	100	14
−45	100	18
−34	100	23
−16	100	32
−1	93	32
0	93	36
2	93	41
4	93	45
10	86	45
15	86	50
19	86	55
27	86	59
36	86	64
43	86	68
59	86	73
81	86	77
92	79	77
101	71	77
103	64	77
107	64	82
108	57	82
132	50	82
150	43	82
187	36	82
195	29	82
215	29	86
220	21	86
251	21	91
275	14	91
326	7	91
336	7	95
432	0	95
Complete group
−123	100	4
−108	100	8
−102	100	12
−99	100	15
−52	100	19
−45	100	23
−34	100	27
−16	100	35
−6	95	35
−1	91	35
0	91	38
1	91	46
2	86	46
4	86	50
10	82	50
15	82	54
19	82	58
22	82	62
27	77	65
36	77	69
43	77	73
59	77	77
81	77	81
83	73	81
90	68	81
92	64	81
101	59	81
103	55	81
107	55	85
108	45	85
132	41	85
150	36	85
187	32	85
195	27	85
201	27	88
215	23	88
220	18	88
251	18	92
264	14	92
275	9	92
326	5	92
336	5	96
432	0	96

The resulting sensitivity and specificity values are presented for the different cut-off numbers for the Validated, Quality, and Complete groups. The value to determine the call of whether the person committed the mock crime (i.e., the decision number) represents the difference in the number of activated voxels for the Mock-crime minus neutral contrast subtracted by the number of activated voxels for the envelope minus neutral contrast. If the decision number is less than the cut-off value, then the call of no sabotage is made. Conversely, if the decision number is greater than the cut-off value, then a call of committing the sabotage is made. The value of “0” was used in the a priori hypothesis in this study and is shown in bold. As can be seen from the above table, small changes in cut-off values can improve the specificity without much decrement to the sensitivity.