Numerical　induction　and　letter　induction　are　two　kinds　of　important　subtypes　of　induction.　Analyzing　their　neural　correlations　is　very　important　for　understanding　the　common　mechanism　of　induction.　Previous　comparative　studies　on　number　cognition　and　letter　comprehension　were　mainly　based　on　a　group　of　comparative　experiment　designs　and　their　neuroimaging　data.　However,　because　of　the　many-to-many　structure-function　relationships,　it　is　difficult　to　understand　neural　correlations　between　number　cognition　and　letter　comprehension,　especially　in　complex　cognitive　functions,　such　as　induction,　only　based　on　single-task　or　few-task　neuroimaging　data　within　an　experimental　lab.　This　paper　proposes　a　systematic　approach　to　analyze　the　similarity　and　disimilarity　of　neural　pattern　between　numerical　and　letter　induction　by　using　Data-Brain　driven　integration　evidence.　Under　the　four　dimensions　of　Data-Brain,　a　group　of　internal　and　external　evidence　is　collected.　A　three　stages　multi-task　analytical　method　is　proposed　to　understand　the　similarity　and　disimilarity　of　neural　pattern　between　numerical　and　letter　induction,　by　combining　meta-analysis　and　representational　similarity.　Results　show　that　more　activation　specific　for　inductive　reasoning　is　left　MFG　and　IFG.　And　number　inductive　reasoning　and　letter　inductive　reasoning　have　high　neural　pattern　similarity　in　the　IFG　and　MFG,　and　a　significant　main　effect　of　inductive　reasoning　is　in　the　left　MFG.　Other　hand,　the　method　can　supplementary　proof　some　results,　it　has　important　implications　for　understand　the　brain　mechanism　of　information　processing.