AbstractBackground and purposeCurrently, the extent of cholinergic basal forebrain atrophy in relatively pure limbic TAR DNA‐binding protein 43 (TDP‐43) pathology compared with relatively pure Alzheimer disease (AD) is unclear.MethodsWe compared antemortem magnetic resonance imaging (MRI)‐based atrophy of the basal forebrain and medial and lateral temporal lobe volumes between 10 autopsy cases with limbic TDP‐43 pathology and 33 cases with AD pathology on postmortem neuropathologic examination from the Alzheimer's Disease Neuroimaging Initiative cohort. For reference, we studied MRI volumes from cognitively healthy, amyloid positron emission tomography‐negative subjects (n = 145). Group differences were assessed using Bayesian analysis of covariance. In addition, we assessed brain‐wide regional volume changes using partial least squares regression (PLSR).ResultsWe found extreme evidence (Bayes factor [BF]01 > 600) for a smaller basal forebrain volume in both TDP‐43 and AD cases compared with amyloid‐negative controls, and moderate evidence (BF01 = 4.9) that basal forebrain volume was not larger in TDP‐43 than in AD cases. The ratio of hippocampus to lateral temporal lobe volumes discriminated between TDP‐43 and AD cases with an accuracy of 0.78. PLSR showed higher gray matter in lateral temporal lobes and cingulate and precuneus, and reduced gray matter in precentral and postcentral gyri and hippocampus in TDP‐43 compared with AD cases.ConclusionsAtrophy of the cholinergic basal forebrain appears to be similarly pronounced in cases with limbic TDP‐43 pathology as in AD. This suggests that a clinical trial of the efficacy of cholinesterase inhibitors in amyloid‐negative cases with amnestic dementia and an imaging signature of TDP‐43 pathology may be warranted.