Premium
O1‐12‐03: Iterative Deconvolution Method for Enhanced Quantitative Accuracy of PET/CT Imaging of Alzheimer's Disease Plaques in Transgenic Mice
Author(s) -
Slavine Nikolai,
McColl Roderick,
Kulkarni Padmakar V.
Publication year - 2016
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2016.06.360
Subject(s) - deconvolution , iterative reconstruction , image quality , pet imaging , nuclear medicine , image noise , image resolution , pet ct , noise (video) , computer science , artificial intelligence , biomedical engineering , medicine , positron emission tomography , algorithm , image (mathematics)
SMCA has an extra penalty on canonical variables of imaging and genetic features, and take two graphs, representing the intra-class and inter-class connections respectively, to encourage the closeness between subjects in the same diagnosis groups and distance between subjects from different diagnosis groups. Data was downloaded from Alzheimer’s Disease Neuroimaging Initiative (ADNI)(adni.loni.usc.edu). 116 ROI measures of amyloid imaging were obtained following the pipeline in [1]. Genotype data of 7517 SNPs in 22 AD risk genes [2] (boundary:100kb) were extracted based on quality controlled and imputed data. There were 977 participants with both amyloid and genotype data (230 cognitively normal controls, 91 with subjective memory concerns, 289 early MCI, 192 late MCI and 175 AD). With the regression weights derived from the CN subjects, amyloid imaging measures were pre-adjusted for baseline age, gender, and education. Results: We compared SMCA with SCCA and 10-fold cross-validation was performed to remove potential bias. SMCA and SCCA shared the same fold partition. In the results, SMCA consistently outperforms SCCAwith significantly higher correlation coefficients between canonical variables in both training and test datasets(Table.1). SCCA identifiesw2000 SNPs jointly affecting w15 brain regions, whereas SMCA only identified rs429358 in APOE with both discriminative power and contribution to the amyloid changes inw20 brain regions such as superior frontal gyri, middle frontal gyri, middle occipital gyri, etc. Conclusions: We proposed and applied a new model SMCA to explore discriminative amyloid imaging and genetic associations, which has demonstrated consistent superiority over traditional SCCA. Further work is indicated toward optimized methods for genome wide search for amyloid and diagnostic markers in AD. References [1]Bioinformatics 30(17),564-571,2014. [2]Nature genetics 45(12),14521458,2014.