AbstractBackgroundA better understanding of drivers behind neuropsychiatric symptoms (NPS) in neurocognitive diseases, such as Alzheimer’s disease, can potentially guide drug development. Cross‐sectional studies have related apathy, depression and anxiety to beta‐amyloid‐ (Aβ) and tau‐pathology, neurodegeneration as well as cognitive deficits already at preclinical stages. However, longitudinal studies are scarce and often small. We investigated effects of neuropathology and cognition on the longitudinal trajectories of NPS in a large group of cognitively unimpaired subjects (CU).MethodThe sample originated from the Swedish BioFINDER study. CU subjects with at least one NPS assessment during the biennial follow‐up of maximum 8 years were included (n=356). The Apathy Evaluation Scale (AES) and the Hospital Depression and Anxiety Scale (HADS) assessed apathy as well as anxious and depressive symptoms, respectively. The Mini Mental State Examination (MMSE) and a modified Preclinical Alzheimer Cognitive Composite (mPACC) measured cognition longitudinally. Plasma P‐tau217, CSF Aβ42/Aβ40 ratio and CSF neurofilament light (NfL) were quantified at baseline using immunoassays. Magnetic Resonance Imaging quantified white matter lesions (WML). Subject‐specific cognitive change per year (slope) was obtained by linear regression models. Linear mixed effect models were used to test associations between longitudinal NPS and other measures (including longitudinal cognition).ResultsSubjects with lower levels of CSF Aβ42/40 (β=‐0.060, p=0.01) or higher plasma P‐tau217 (β=0.253, p=0.015) at study start demonstrated accelerated levels of apathy over time. On a trend level, a high load of WML predicted over time high levels of apathy (β=0.183, p=0.053). Increased levels of longitudinal anxiety were predicted by the interaction between time and lower CSF Aβ42/40 (β=‐0.283, p=0.005) or higher CSF NfL (β=0.054, p=0.024). Furthermore, more rapid decline in mPACC was associated with more rapid progression of apathy (β=‐0.478, p=<0.001) and anxiety (β=‐0.066, p=<0.023). Similar results were obtained for longitudinal associations between MMSE and apathy (β=‐0.673, p=<0.001).ConclusionThese results suggests that the longitudinal trajectories of apathy and anxiety already in CU subjects are related to key hallmarks of Alzheimer´s disease, including Aβ‐ or tau‐pathology at baseline, and further are parallel with cognitive decline over time.