Neuroimgaing processing library

Requirements

• Downloaded SPM toolbox

Processing data

The analysis pipeline can be described as following: Rearranging folders -> Processing images -> Extracting features -> Apply machine learning algorithms for classification See more details of implementation in the wiki. In order to start analysis Main_script both for Matlab and Python has to be used.

Results for current repopsitory state

AV45 images Processing results

	Total	Mean Age		Std Age		MMSE
		Male	Female	Male	Female	Male	Female
Normal	408	74.13	72.26	6.08	5.68	28.67	28.96
MCI	87	75.2	75.32	7.16	5.2	27.59	27.48
AD	237	75.46	72.75	7.45	7.65	21.23	21.09

Test dataset size: 74, where: Normal: 39; MCI: 9; AD: 26

Confusion Matrix	Normilized CM
Normal MCI AD Normal 37 2 0 MCI 0 8 1 AD 1 2 23	Normal MCI AD Normal 0.95 0.05 0 MCI 0 0.88 0.11 AD 0.039 0.08 0.88

Best mean accuracy for 10 folds: 0.93

---

FDG images Processing results

	Total	Mean Age		Std Age		MMSE
		Male	Female	Male	Female	Male	Female
Normal	238	74.11	72.04	5.9	6.05	28.7	29.02
MCI	57	76.23	75.24	6.9	5.54	27.5	27.46
AD	202	75.5	72.6	7.63	7.66	21.04	21.06

Test dataset size: 55, where: Normal: 34; MCI: 7; AD: 14

Confusion Matrix	Normilized CM
Normal MCI AD Normal 34 0 0 MCI 0 7 0 AD 0 0 14	Normal MCI AD Normal 1 0 0 MCI 0 1 0 AD 0 0 1

Best mean accuracy for 10 folds: 0.99