Premium
Connectivity‐based differentiation of agrammatic and logopenic variants of primary progressive aphasia using a whole brain data‐driven approach
Author(s) -
Jenkins Lisanne M,
Behn Jordan Quinn,
Khan Raiyan R,
Wang Lei,
Rogalski Emily J,
Greicius Michael D,
Mesulam Marsel,
Bonakdarpour Borna
Publication year - 2020
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.043956
Subject(s) - primary progressive aphasia , post hoc analysis , post hoc , voxel , aphasia , analysis of variance , psychology , medicine , audiology , frontotemporal dementia , neuroscience , computer science , artificial intelligence , disease , pathology , dementia
Background Primary progressive aphasia (PPA) is a syndrome of progressive language impairment caused by neurodegenerative disease. We previously showed using seed‐based connectivity analysis that connectivity between three language network regions is decreased in PPA. Furthermore, we showed selective connectivity disruptions corresponding to grammar, repetition, naming and comprehension deficits. The current investigation used a data driven approach to uncover signatures of resting state connectivity disruptions differentiating agrammatic PPA subtypes (PPA‐G) from logopenic PPA subtypes (PPA‐L). Method One hundred and six individuals participated in this study, for which 73 PPA patients were compared to 33 healthy controls (Table 1). Diagnosis of PPA subtype was made by clinical consensus using established criteria. Resting state functional MRI (rsfMRI) was collected using a Siemens 3T scanner over 10 minutes with eyes open. Data were analyzed using FSL tools. A group independent components analysis generated 30 independent components (ICs). Participant’s individual time series and component spatial maps were estimated using dual regression. Voxel‐wise non‐parametric permutation testing (n=5000) tested for significant differences in connectivity within each IC across groups. Follow‐up one‐way ANOVAs with Tukey post‐hoc tests were conducted. Result Post hoc tests found PPA subgroup(s) had reduced connectivity than controls in three FSL‐identified ICs showing significant group differences (Figure 1). PPA‐L had significantly lower connectivity within IC1 in the cuneus. PPA‐G had significantly lower connectivity within IC14 in Broca’s area; and all three PPA subtypes had significantly lower connectivity within IC25 in the left middle temporal gyrus (MTG). Conclusion We identified three components in which PPA subgroup(s) had lower connectivity than controls. In PPA‐G, but not PPA‐L, reduced connectivity was within Broca’s area, consistent with this region’s importance for fluency and grammar. In PPA‐L, but not PPA‐G, reduced connectivity was within the cuneus. This finding may seem surprising, however, it may reflect higher incidence in PPA‐L of Alzheimer pathology, which is known to have the precuneus as a major target of degeneration. All PPA subtypes had reduced connectivity within the posterior MTG, reflecting the common occurrence of atrophy in this area.