Network　analysis　of　human　brain　connectivity　based　on　graph　theory　has　consistently　identified　sets　of　regions　that　are　critically　important　for　enabling　efficient　information　integration　and　communication,　especially　for　the　understanding　of　cognitive　functions,　the　discoveries　of　aging　effects　and　the　network　change　due　to　brain　diseases.　Two　major　approaches,　hub　measurement　(HM)　and　vulnerability　measurement　(VM),　have　been　proposed　to　detect　these　＇important　nodes＇　within　brain　network　organization.　However,　the　relationship　between　the　spatial　localization　and　the　number　of　these　identified　nodes　found　using　HM　and　VM　approaches　respectively　is　still　unknown.　In　this　study,　we　aim　to　figure　out　the　relationships　between　the　identified　critical　nodes　of　brain　network　based　on　various　HM　and　VM　methods　with　DTI-based　structural　brain　network.　Two　factors　of　parcellation　atlases　and　level　of　scale　are　also　considered　to　address　the　effects　in　the　definition　of　these　nodes.　From　the　results,　the　great　consistency　is　existed　between　the　node　identification　using　HM　and　VM　approaches　in　the　same　atlases,　but　the　divergence　between　different　atlases　and　level　of　node　scale.