GAT-LI: a graph attention network based learning and interpreting method for functional brain network classification

Table 7 Analyses of 10 connections in synthetic graph dataset

Connection ID	Node ID	Class_one mean connection	Class_two mean connection	Mean difference	p value
1	16	0.5052	0.4916	0.0135	0.0362**
1	5	0.5052	0.4916	0.0135	0.0362**
2	13	0.5160	0.4962	0.0197	0.0023**
2	5	0.5160	0.4962	0.0197	0.0023**
3	4	0.5079	0.4904	0.0175	0.0073**
3	1	0.5079	0.4904	0.0175	0.0073**
4	15	0.5061	0.4995	0.0066	0.3090
4	5	0.5061	0.4995	0.0066	0.3090
5	22	0.5068	0.4978	0.0091	0.1554
5	5	0.5068	0.4978	0.0091	0.1554
6	27	0.5144	0.4864	0.0272	0.00002**
6	30	0.5144	0.4864	0.0272	0.00002**
7	14	0.5109	0.4837	0.0272	0.00002**
7	5	0.5109	0.4837	0.0272	0.00002**
8	28	0.5121	0.4860	0.0260	0.0039**
8	28	0.5121	0.4860	0.0260	0.0039**
9	27	0.5064	0.4900	0.0164	0.0109**
9	1	0.5064	0.4900	0.0164	0.0109**
10	26	0.5059	0.4846	0.0219	0.0011**
10	19	0.5059	0.4846	0.0219	0.0011**

ISSN: 1471-2105