Premium
The link between brain functional network connectivity and genetic risk of Alzheimer's disease
Author(s) -
Sendi Mohammad S.E.,
Zendehrouh Elaheh,
Miller Robyn L.,
Mormino Elizabeth C.,
Salat David H.,
Calhoun Vince D.
Publication year - 2021
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 - 10.1002/alz.050101
Subject(s) - allele , correlation , medicine , genetics , psychology , biology , gene , mathematics , geometry
Background Individuals carrying the ɛ4 allele have the highest risk of Alzheimer’s disease (AD) compared with those carrying ɛ3 and ɛ2 allele, whereas ɛ2 allele has the lowest risk. Although previous studies explored the link between the genetic risk of AD and static functional network connectivity (sFNC), limited studies have evaluated the association between dynamic FNC (dFNC) and AD risk. Here, we explore how the dFNC differs between individuals with genetic risk for AD. Method We used rs‐fMRI data of 991 healthy brains (404 females) and their demographic information from the Open Access Series of Imaging Studies‐3 cohort. The participants' age at scanning time was ranging from 42.46 to 95.39, with a mean of 69.81. We put the data into three groups including group1 (N=135, 63 females) including subjects with ɛ2 allele (i.e., ɛ2/ ɛ2, ɛ2/ ɛ3, and ɛ3/ɛ2), group2 (N=558, 219 females) including subjects with only ɛ3 allele (i.e., ɛ3/ ɛ3), and group3 (N=298, 122 females) including subjects with ɛ4 allele (i.e., ɛ3/ ɛ4, ɛ4/ ɛ3, and ɛ4/ɛ4). Age and gender were not significantly different across groups. Group‐ICA was used to extract 53 components. The sliding window and Pearson correlation were used to measure the dFNC among components K‐means algorithm was applied to partition dFNC windows into three distinct states. We calculated each subject's occupancy rate (OCR) in each state. A two‐sample t‐test was used to compare the OCR of groups in each state (Fig. 1). Result Subject with a lower AD risk spend more time in state1 with more positive connectivity within cognitive control network (CCN) and between CCN and sensory network (Fig. 2A and Fig. 2B: p corrected <0.05). Interestingly, in this state, the difference of OCR among subjects with different AD risk was more significant in females (Fig. 2C), while males did not show any significant difference in their OCR across three groups (Fig. 2D). Females with higher AD risk had more OCR in state 3 with relatively lower within CCN connectivity. Conclusion Results support the use of dFNC features as a potential biomarker of AD genetic risk for females. We also showed the role of CCN connectivity associated with the AD risk.